<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Classified</title>
    <description>The latest articles on DEV Community by Classified (@iamclassified).</description>
    <link>https://dev.to/iamclassified</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4143571%2F87426864-f60a-4480-b3e5-a0b8c21aab5c.jpg</url>
      <title>DEV Community: Classified</title>
      <link>https://dev.to/iamclassified</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/iamclassified"/>
    <language>en</language>
    <item>
      <title>#ai</title>
      <dc:creator>Classified</dc:creator>
      <pubDate>Thu, 29 Oct 2026 05:52:56 +0000</pubDate>
      <link>https://dev.to/iamclassified/ai-234m</link>
      <guid>https://dev.to/iamclassified/ai-234m</guid>
      <description>&lt;div class="ltag__link--embedded"&gt;
  &lt;div class="crayons-story "&gt;
  &lt;a href="https://dev.to/iamclassified/hunter-signal-lab-i-built-a-lazy-morning-brief-so-i-could-sleep-instead-of-staring-at-charts-2c01" class="crayons-story__hidden-navigation-link"&gt;Hunter Signal Lab: I Built a Lazy Morning Brief So I Could Sleep Instead of Staring at Charts&lt;/a&gt;


  &lt;div class="crayons-story__body crayons-story__body-full_post"&gt;
      &lt;a href="https://dev.to/iamclassified/hunter-signal-lab-i-built-a-lazy-morning-brief-so-i-could-sleep-instead-of-staring-at-charts-2c01" class="crayons-article__context-note crayons-article__context-note__feed"&gt;&lt;p&gt;Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady Submission 🦥&lt;/p&gt;

&lt;/a&gt;
    &lt;div class="crayons-story__top"&gt;
      &lt;div class="crayons-story__meta"&gt;
        &lt;div class="crayons-story__author-pic"&gt;

          &lt;a href="/iamclassified" class="crayons-avatar  crayons-avatar--l  "&gt;
            &lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4143571%2F87426864-f60a-4480-b3e5-a0b8c21aab5c.jpg" alt="iamclassified profile" class="crayons-avatar__image" width="460" height="460"&gt;
          &lt;/a&gt;
        &lt;/div&gt;
        &lt;div&gt;
          &lt;div&gt;
            &lt;a href="/iamclassified" class="crayons-story__secondary fw-medium m:hidden"&gt;
              Classified
            &lt;/a&gt;
            &lt;div class="profile-preview-card relative mb-4 s:mb-0 fw-medium hidden m:inline-block"&gt;
              
                Classified
                
                
              
              &lt;div id="story-author-preview-content-4918813" class="profile-preview-card__content crayons-dropdown branded-7 p-4 pt-0"&gt;
                &lt;div class="gap-4 grid"&gt;
                  &lt;div class="-mt-4"&gt;
                    &lt;a href="/iamclassified" class="flex"&gt;
                      &lt;span class="crayons-avatar crayons-avatar--xl mr-2 shrink-0"&gt;
                        &lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4143571%2F87426864-f60a-4480-b3e5-a0b8c21aab5c.jpg" class="crayons-avatar__image" alt="" width="460" height="460"&gt;
                      &lt;/span&gt;
                      &lt;span class="crayons-link crayons-subtitle-2 mt-5"&gt;Classified&lt;/span&gt;
                    &lt;/a&gt;
                  &lt;/div&gt;
                  &lt;div class="print-hidden"&gt;
                    
                      Follow
                    
                  &lt;/div&gt;
                  &lt;div class="author-preview-metadata-container"&gt;&lt;/div&gt;
                &lt;/div&gt;
              &lt;/div&gt;
            &lt;/div&gt;

          &lt;/div&gt;
          &lt;a href="https://dev.to/iamclassified/hunter-signal-lab-i-built-a-lazy-morning-brief-so-i-could-sleep-instead-of-staring-at-charts-2c01" class="crayons-story__tertiary fs-xs"&gt;&lt;time&gt;Oct 26&lt;/time&gt;&lt;span class="time-ago-indicator-initial-placeholder"&gt;&lt;/span&gt;&lt;/a&gt;
        &lt;/div&gt;
      &lt;/div&gt;

    &lt;/div&gt;

    &lt;div class="crayons-story__indention"&gt;
      &lt;h2 class="crayons-story__title crayons-story__title-full_post"&gt;
        &lt;a href="https://dev.to/iamclassified/hunter-signal-lab-i-built-a-lazy-morning-brief-so-i-could-sleep-instead-of-staring-at-charts-2c01" id="article-link-4918813"&gt;
          Hunter Signal Lab: I Built a Lazy Morning Brief So I Could Sleep Instead of Staring at Charts
        &lt;/a&gt;
      &lt;/h2&gt;
        &lt;div class="crayons-story__tags"&gt;
            &lt;a class="crayons-tag  crayons-tag--monochrome " href="/t/devchallenge"&gt;&lt;span class="crayons-tag__prefix"&gt;#&lt;/span&gt;devchallenge&lt;/a&gt;
            &lt;a class="crayons-tag  crayons-tag--monochrome " href="/t/hf26challenge"&gt;&lt;span class="crayons-tag__prefix"&gt;#&lt;/span&gt;hf26challenge&lt;/a&gt;
            &lt;a class="crayons-tag  crayons-tag--monochrome " href="/t/opensource"&gt;&lt;span class="crayons-tag__prefix"&gt;#&lt;/span&gt;opensource&lt;/a&gt;
            &lt;a class="crayons-tag  crayons-tag--monochrome " href="/t/ai"&gt;&lt;span class="crayons-tag__prefix"&gt;#&lt;/span&gt;ai&lt;/a&gt;
        &lt;/div&gt;
      &lt;div class="crayons-story__bottom"&gt;
        &lt;div class="crayons-story__details"&gt;
          &lt;a href="https://dev.to/iamclassified/hunter-signal-lab-i-built-a-lazy-morning-brief-so-i-could-sleep-instead-of-staring-at-charts-2c01" class="crayons-btn crayons-btn--s crayons-btn--ghost crayons-btn--icon-left"&gt;
            &lt;div class="multiple_reactions_aggregate"&gt;
              &lt;span class="multiple_reactions_icons_container"&gt;
                  &lt;span class="crayons_icon_container"&gt;
                    &lt;img src="https://assets.dev.to/assets/raised-hands-74b2099fd66a39f2d7eed9305ee0f4553df0eb7b4f11b01b6b1b499973048fe5.svg" width="24" height="24"&gt;
                  &lt;/span&gt;
                  &lt;span class="crayons_icon_container"&gt;
                    &lt;img src="https://assets.dev.to/assets/exploding-head-daceb38d627e6ae9b730f36a1e390fca556a4289d5a41abb2c35068ad3e2c4b5.svg" width="24" height="24"&gt;
                  &lt;/span&gt;
                  &lt;span class="crayons_icon_container"&gt;
                    &lt;img src="https://assets.dev.to/assets/multi-unicorn-b44d6f8c23cdd00964192bedc38af3e82463978aa611b4365bd33a0f1f4f3e97.svg" width="24" height="24"&gt;
                  &lt;/span&gt;
              &lt;/span&gt;
              &lt;span class="aggregate_reactions_counter"&gt;14&lt;span class="hidden s:inline"&gt;&amp;nbsp;reactions&lt;/span&gt;&lt;/span&gt;
            &lt;/div&gt;
          &lt;/a&gt;
            &lt;a href="https://dev.to/iamclassified/hunter-signal-lab-i-built-a-lazy-morning-brief-so-i-could-sleep-instead-of-staring-at-charts-2c01#comments" class="crayons-btn crayons-btn--s crayons-btn--ghost crayons-btn--icon-left flex items-center"&gt;
              

              &lt;span class="hidden s:inline"&gt;Add&amp;nbsp;Comment&lt;/span&gt;
            &lt;/a&gt;
        &lt;/div&gt;
        &lt;div class="crayons-story__save"&gt;
          &lt;small class="crayons-story__tertiary fs-xs mr-2"&gt;
            8 min read
          &lt;/small&gt;
        &lt;/div&gt;
      &lt;/div&gt;
    &lt;/div&gt;
  &lt;/div&gt;
&lt;/div&gt;

&lt;/div&gt;


</description>
      <category>sideprojects</category>
      <category>automation</category>
      <category>productivity</category>
      <category>ai</category>
    </item>
    <item>
      <title>Hunter Signal Lab: I Built a Lazy Morning Brief So I Could Sleep Instead of Staring at Charts</title>
      <dc:creator>Classified</dc:creator>
      <pubDate>Mon, 26 Oct 2026 03:50:18 +0000</pubDate>
      <link>https://dev.to/iamclassified/hunter-signal-lab-i-built-a-lazy-morning-brief-so-i-could-sleep-instead-of-staring-at-charts-2c01</link>
      <guid>https://dev.to/iamclassified/hunter-signal-lab-i-built-a-lazy-morning-brief-so-i-could-sleep-instead-of-staring-at-charts-2c01</guid>
      <description>&lt;p&gt;****&lt;em&gt;This is a submission for the &lt;a href="https://dev.to/challenges/hacktoberfest-week3-2026-10-19"&gt;Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady&lt;/a&gt;&lt;/em&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What I Built
&lt;/h2&gt;

&lt;p&gt;I am a student. I also trade and analyse the crypto market while preparing for my studies. Between lectures, revision, chart analysis, sleep and everything else, my day had no room left.&lt;/p&gt;

&lt;p&gt;The worst part was the charts. Every day I sat and looked for &lt;strong&gt;liquidity events&lt;/strong&gt;: moments when price sweeps a level where many traders have their orders, then breaks out or reverses. Finding them by eye takes hours.&lt;/p&gt;

&lt;p&gt;So I built &lt;strong&gt;Hunter Signal Lab&lt;/strong&gt;. It watches real crypto candles for me, finds liquidity events using fixed written rules, draws a small chart snapshot of each one, and writes a short plain-English brief. Now I wake up, open the dashboard, press &lt;strong&gt;Brief&lt;/strong&gt;, and read what happened overnight. The time I used to spend analysing, I now spend sleeping.&lt;/p&gt;

&lt;p&gt;Who gets to be lazier because of it? Any student or part-time trader who cannot sit in front of charts all day.&lt;/p&gt;

&lt;p&gt;One important thing: &lt;strong&gt;Hunter does not predict anything and gives no trading advice.&lt;/strong&gt; It only records what happened and explains it.&lt;/p&gt;

&lt;p&gt;![Dashboard overview](&lt;br&gt;
&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Flq04xx60c83uce2g8lcz.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Flq04xx60c83uce2g8lcz.png" alt="Hunter Signal Lab dashboard showing the liquidity brief with a chart snapshot for each event" width="799" height="385"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;)&lt;/p&gt;

&lt;p&gt;![Dashboard overview](&lt;br&gt;
&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F3fqdbjh2utygkl2ooqtv.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F3fqdbjh2utygkl2ooqtv.png" alt="Hunter Signal Lab dashboard showing the liquidity brief with a chart snapshot for each event showing watchlist" width="800" height="373"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;)&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F2o6ogyjfy4su76l4sxbk.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F2o6ogyjfy4su76l4sxbk.png" alt="The liquidity brief with chart snapshots" width="800" height="381"&gt;&lt;/a&gt;&lt;/p&gt;
&lt;h2&gt;
  
  
  Demo
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Live app (on Render):&lt;/strong&gt; &lt;a href="https://hunter-signal-lab-web.onrender.com" rel="noopener noreferrer"&gt;https://hunter-signal-lab-web.onrender.com&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;API docs:&lt;/strong&gt; &lt;a href="https://hunter-signal-lab-api.onrender.com/docs" rel="noopener noreferrer"&gt;https://hunter-signal-lab-api.onrender.com/docs&lt;/a&gt;
&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Free hosting sleeps when idle, so the first load can take 30 to 60 seconds. Please be patient with it. It is a sloth too. 🦥&lt;/p&gt;

&lt;p&gt;![Top 5 events and watchlist](&lt;br&gt;
&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fgd6dv45ksk0cl4jsfzrp.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fgd6dv45ksk0cl4jsfzrp.png" alt="over view showing brief result" width="799" height="476"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;)&lt;/p&gt;

&lt;p&gt;![Top Events No.1]&lt;br&gt;
&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F2o6ogyjfy4su76l4sxbk.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F2o6ogyjfy4su76l4sxbk.png" alt="Top Event" width="800" height="381"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;![Top Events No.2]&lt;br&gt;
&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F6572poyik7vnomtxk86t.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F6572poyik7vnomtxk86t.png" alt="Top Event" width="799" height="403"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;![Top Event No.3]&lt;br&gt;
&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F81furpz318151252zd2n.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F81furpz318151252zd2n.png" alt="Top Event No.3" width="800" height="373"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;![Event explorer](&lt;br&gt;
&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fqg67gpg73mvsfg0ddthr.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fqg67gpg73mvsfg0ddthr.png" alt="Event explorer" width="800" height="975"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;)&lt;/p&gt;
&lt;h2&gt;
  
  
  Code
&lt;/h2&gt;


&lt;div class="ltag-github-readme-tag"&gt;
  &lt;div class="readme-overview"&gt;
    &lt;h2&gt;
      &lt;img src="https://assets.dev.to/assets/github-logo-5a155e1f9a670af7944dd5e12375bc76ed542ea80224905ecaf878b9157cdefc.svg" alt="GitHub logo"&gt;
      &lt;a href="https://github.com/IamClassified9" rel="noopener noreferrer"&gt;
        IamClassified9
      &lt;/a&gt; / &lt;a href="https://github.com/IamClassified9/hunter-signal-lab" rel="noopener noreferrer"&gt;
        hunter-signal-lab
      &lt;/a&gt;
    &lt;/h2&gt;
    &lt;h3&gt;
      
    &lt;/h3&gt;
  &lt;/div&gt;
  &lt;div class="ltag-github-body"&gt;
    
&lt;div id="readme" class="md"&gt;&lt;div class="markdown-heading"&gt;
&lt;h1 class="heading-element"&gt;Hunter Signal Lab&lt;/h1&gt;
&lt;/div&gt;
&lt;p&gt;Hunter Signal Lab detects &lt;strong&gt;liquidity events&lt;/strong&gt;—sweeps, grabs, breakouts, and failed breakouts—from real crypto candles and displays the evidence on an interactive chart.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Current scope: Phase 1 — event detection and evidence review.&lt;/strong&gt; The system records and explains observed market events; it does not predict future prices or generate trade signals.&lt;/p&gt;
&lt;div class="markdown-heading"&gt;
&lt;h2 class="heading-element"&gt;Live deployment&lt;/h2&gt;
&lt;/div&gt;
&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Live dashboard:&lt;/strong&gt; &lt;a href="https://hunter-signal-lab-web.onrender.com" rel="nofollow noopener noreferrer"&gt;https://hunter-signal-lab-web.onrender.com&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Backend API:&lt;/strong&gt; &lt;a href="https://hunter-signal-lab-api.onrender.com/" rel="nofollow noopener noreferrer"&gt;https://hunter-signal-lab-api.onrender.com/&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Interactive API documentation:&lt;/strong&gt; &lt;a href="https://hunter-signal-lab-api.onrender.com/docs" rel="nofollow noopener noreferrer"&gt;https://hunter-signal-lab-api.onrender.com/docs&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Source code:&lt;/strong&gt; &lt;a href="https://github.com/IamClassified9/hunter-signal-lab" rel="noopener noreferrer"&gt;https://github.com/IamClassified9/hunter-signal-lab&lt;/a&gt;
&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;The services use free hosting, which may sleep while idle. The first request after inactivity can take 30–60 seconds.&lt;/p&gt;
&lt;div class="markdown-heading"&gt;
&lt;h2 class="heading-element"&gt;What it does&lt;/h2&gt;
&lt;/div&gt;
&lt;ul&gt;
&lt;li&gt;Fetches closed OHLCV candles from a public exchange through CCXT; no exchange API keys are required for public market data.&lt;/li&gt;
&lt;li&gt;Identifies swing highs and lows, plus equal-high/equal-low liquidity levels.&lt;/li&gt;
&lt;li&gt;Detects five event types using explicit rules and information available at the time, avoiding look-ahead in event detection.&lt;/li&gt;
&lt;li&gt;Exposes market data and analysis through a FastAPI backend and a React…&lt;/li&gt;
&lt;/ul&gt;&lt;/div&gt;
  &lt;/div&gt;
  &lt;div class="gh-btn-container"&gt;&lt;a class="gh-btn" href="https://github.com/IamClassified9/hunter-signal-lab" rel="noopener noreferrer"&gt;View on GitHub&lt;/a&gt;&lt;/div&gt;
&lt;/div&gt;


&lt;p&gt;The repo has a full README with the event rules, API table, configuration and known limitations. It also has 150 backend tests. The exchange is mocked, so they need no network.&lt;/p&gt;

&lt;p&gt;![150 backend tests passing](&lt;br&gt;
&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fl1wcuhidt00yzdi6pcrz.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fl1wcuhidt00yzdi6pcrz.png" alt="150 backend tests" width="800" height="538"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;)&lt;/p&gt;

&lt;h2&gt;
  
  
  How I Built It
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;The engine (no AI needed).&lt;/strong&gt; Hunter fetches closed candles through CCXT, finds swing highs and lows, and finds clusters of equal highs and lows. It detects five event types with explicit rules: breakout, failed breakout, swing high or low sweep, and liquidity grab. It uses only information that was available at the time (no look-ahead). Tests prove this by comparing a run on partial data with the full run.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;The stack.&lt;/strong&gt; FastAPI backend, React and TypeScript dashboard, a watchlist saved in the browser, a top-5 event list, and JSON export.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Chart snapshots.&lt;/strong&gt; For each event in the brief, the backend draws a small SVG chart with the event price and the linked level marked. Nothing is sent to an outside chart service.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;The AI part.&lt;/strong&gt; The brief is written by Gemma. The rule is strict: Hunter supplies the facts, Gemma only explains them. Gemma gets no permission to predict or advise, and the engine's numbers are shown beside every sentence so I can check it.&lt;/p&gt;

&lt;h3&gt;
  
  
  The messy part: how I got to Gemma
&lt;/h3&gt;

&lt;p&gt;This took me about four days of going back and forth. I want to tell it honestly.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;1. TabPFN.&lt;/strong&gt; My first plan was to add TabPFN as a prediction engine. It is built for tables of data, which fits market data. But it needs more resources than my laptop has, and I could not run it. I also realised that predicting anything honestly needs a measured history of what price did after each event, and I did not have that yet. So the dashboard panel says "Not connected" on purpose, and it sits on my roadmap behind proper outcome testing.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;2. Qwen 3 through Kaggle.&lt;/strong&gt; I wanted an open model to write the brief. Again My laptop was struggling, so I tried to reach Qwen 3 through Kaggle's resources. It became complicated, it kept not finishing, and I was burning days.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;3. Gemma on my own machine with Ollama.&lt;/strong&gt; Next I tried hosting Gemma locally. Ollama would not run smoothly on my system. It lagged, it did not have the resources for it.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;4. Gemma through a hosted API.&lt;/strong&gt; Finally I used hosted Gemma. A direct test call worked, and I wired it into the backend. The key and model name live only in the backend environment, never in the frontend.&lt;/p&gt;

&lt;h2&gt;
  
  
  Use of Gemma
&lt;/h2&gt;

&lt;p&gt;Gemma is the writer of the whole product. Without it, Hunter is a chart tool. With it, Hunter is a morning brief.&lt;/p&gt;

&lt;h3&gt;
  
  
  Gemma at a glance
&lt;/h3&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;&lt;/th&gt;
&lt;th&gt;&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Model&lt;/td&gt;
&lt;td&gt;&lt;a href="https://dev.toopen-weight"&gt;gemma-4-26b-a4b-it&lt;/a&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Access&lt;/td&gt;
&lt;td&gt;Gemini API, through Google's Gen AI SDK&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Where it runs&lt;/td&gt;
&lt;td&gt;My FastAPI backend on Render&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Config&lt;/td&gt;
&lt;td&gt;
&lt;code&gt;BRIEF_MODEL&lt;/code&gt; in the backend environment. Key on the backend only, never in the frontend&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Used for&lt;/td&gt;
&lt;td&gt;Writing the plain-English brief for the top 3 events&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Health check&lt;/td&gt;
&lt;td&gt;&lt;code&gt;/api/ai/health&lt;/code&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h3&gt;
  
  
  The contract: Hunter gives facts, Gemma gives words
&lt;/h3&gt;

&lt;p&gt;For each selected event, the backend sends Gemma only the engine's recorded facts: symbol, timeframe, event type, event price, linked level, level status, volume ratio, structure, event time and detected time. Gemma sends back short plain-English text. For now i instructed it to give only two lines summary.&lt;/p&gt;

&lt;p&gt;The rules I give Gemma:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Explain what the event structure and the linked level mean.&lt;/li&gt;
&lt;li&gt;Say whether the linked level was swept, unswept, or has conflicting evidence.&lt;/li&gt;
&lt;li&gt;Explain the volume ratio without claiming it proves buying or selling pressure.&lt;/li&gt;
&lt;li&gt;Keep the time the event happened separate from the time Hunter detected it.&lt;/li&gt;
&lt;li&gt;Say when follow-through evidence is not available.&lt;/li&gt;
&lt;li&gt;No predictions. No trading advice.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Temperature is set to 0.2 to keep the model close to the numbers.&lt;/p&gt;

&lt;h3&gt;
  
  
  How I keep Gemma honest
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;The numbers stay visible.&lt;/strong&gt; The engine's facts are shown in a separate evidence panel next to Gemma's text, so a reader can check every sentence.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Bad output is rejected.&lt;/strong&gt; If Gemma returns something unusable, the API returns an error. It does not pass bad text to the user.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;No silent fakes.&lt;/strong&gt; If Gemma is unreachable, the dashboard shows the reason in red. A template fallback exists, but it is off unless I switch it on.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Tests.&lt;/strong&gt; [If true: The 150 backend tests include tests for the brief, using a fake model so they run without a network.]&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  Verify it yourself
&lt;/h3&gt;

&lt;ol&gt;
&lt;li&gt;Open the live app and press &lt;strong&gt;Brief&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Open &lt;a href="https://hunter-signal-lab-api.onrender.com/api/ai/health" rel="noopener noreferrer"&gt;https://hunter-signal-lab-api.onrender.com/api/ai/health&lt;/a&gt; to see the Gemma status.&lt;/li&gt;
&lt;li&gt;In the repo, backend/app/ai_service.py shows the call to Gemma.
&lt;/li&gt;
&lt;/ol&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight python"&gt;&lt;code&gt;
&lt;span class="n"&gt;client&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;genai&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nc"&gt;Client&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;api_key&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="n"&gt;api_key&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;

&lt;span class="n"&gt;response&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;client&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="n"&gt;models&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;generate_content&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;
    &lt;span class="n"&gt;model&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="n"&gt;BRIEF_MODEL&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
    &lt;span class="n"&gt;contents&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="n"&gt;prompt&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
    &lt;span class="n"&gt;config&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="n"&gt;types&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nc"&gt;GenerateContentConfig&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;
        &lt;span class="n"&gt;temperature&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="n"&gt;BRIEF_TEMPERATURE&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
    &lt;span class="p"&gt;),&lt;/span&gt;
&lt;span class="p"&gt;)&lt;/span&gt;

&lt;span class="n"&gt;answer&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;response&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="n"&gt;text&lt;/span&gt;

&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;![ALT: /api/ai/health on the live Render app showing Gemma connected](&lt;br&gt;
&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Ffi7am228ot1npro1we86.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Ffi7am228ot1npro1we86.png" alt="api health status" width="800" height="214"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;)&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F2o6ogyjfy4su76l4sxbk.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F2o6ogyjfy4su76l4sxbk.png" alt="ALT: A brief written by Gemma with chart snapshots and an evidence panel" width="800" height="381"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h3&gt;
  
  
  Why I ended up on Gemma
&lt;/h3&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Option&lt;/th&gt;
&lt;th&gt;What happened&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Qwen 3 through Kaggle&lt;/td&gt;
&lt;td&gt;Complicated, never finished, burned days&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Gemma locally with Ollama&lt;/td&gt;
&lt;td&gt;Would not download and run smoothly; my laptop lagged&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;TabPFN&lt;/td&gt;
&lt;td&gt;Needs more resources than my laptop has; no outcome history yet&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Gemma through the Gemini API&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;&lt;strong&gt;Worked, and it is open-weight&lt;/strong&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h3&gt;
  
  
  What I learned about using a small open model
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;Keep the model name in one setting. Spreading it through the code made every switch painful.&lt;/li&gt;
&lt;li&gt;A small model can sound sure of itself. Showing the numbers beside the words is the fix.&lt;/li&gt;
&lt;li&gt;Connectivity issues, a whole lot of times i hit a couple of 429 502 and 503 errors&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;I also switched the setup so the &lt;strong&gt;template fallback is off by default&lt;/strong&gt;. If Gemma fails, the API returns a clear error and the dashboard shows it in red. I did this because a silent template can pretend the AI wrote something when it did not.&lt;/p&gt;

&lt;h3&gt;
  
  
  What works today
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;** The Brief button works end to end. Gemma writes the explanations, the chart snapshots are attached, and the brief can be exported as a report.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Why Render Was the Right Home for This
&lt;/h2&gt;

&lt;p&gt;Render is where Hunter lives, and it fit a student with no budget and no free evenings.&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;One file builds everything.&lt;/strong&gt; A &lt;code&gt;render.yaml&lt;/code&gt; Blueprint defines two services: a Python web service for the FastAPI backend and a static site for the React frontend. I pushed to GitHub, chose New, Blueprint, and Render set both up.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Region matters.&lt;/strong&gt; Some exchanges block requests from US servers. I put the backend in Frankfurt or so the candle data keeps flowing. This was a real problem, and Render's region choice solved it.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Secrets stay off GitHub.&lt;/strong&gt; The Gemma key and model name are set in Render's environment settings for the backend. Nothing secret is in the repo or the frontend.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Clear wiring.&lt;/strong&gt; The frontend gets the backend address through &lt;code&gt;VITE_API_BASE_URL&lt;/code&gt; at build time. The backend allows the frontend through &lt;code&gt;HUNTER_CORS_ORIGINS&lt;/code&gt;. Two settings, and the two services talk.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Health checks.&lt;/strong&gt; &lt;code&gt;/health&lt;/code&gt; and &lt;code&gt;/api/ai/health&lt;/code&gt; report whether the exchange and Gemma are reachable, so I can see if something is wrong without reading logs.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Free tier, honestly.&lt;/strong&gt; It sleeps when idle, so the first load is slow. For a tool I open once each morning, that is fine.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;![Both services running on Render](&lt;br&gt;
&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fjzzekr1xw9pdxyh9mdbw.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fjzzekr1xw9pdxyh9mdbw.png" alt="Render" width="800" height="464"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;) &lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fmg2i1md6dsecz1efknwo.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fmg2i1md6dsecz1efknwo.png" alt="Render" width="800" height="391"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;![render.yaml Blueprint](&lt;br&gt;
&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fj60t3e0ks1if25nuvplc.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fj60t3e0ks1if25nuvplc.png" alt="yaml" width="799" height="401"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;)&lt;/p&gt;

&lt;h2&gt;
  
  
  Why Does Open Innovation Matter?
&lt;/h2&gt;

&lt;p&gt;Well using myself as an example, My laptop could not run open models locally, and I tried three different ways. That failure is exactly why open models matter to me.&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;I could swap models by changing one setting.&lt;/strong&gt; I moved from Qwen to Gemma without redesigning anything, because the model name sits in one environment variable. A closed API would have locked me to one provider.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;The same open model can come home later.&lt;/strong&gt; I used hosted Gemma because my hardware is weak today. Because the weights are open, the same Gemma can run on my own machine once I have a better one. My trading notes would then stay private on my laptop and work offline.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;I could inspect and control it.&lt;/strong&gt; With open tooling I could pin the temperature low and keep the model tied to the numbers it is given, instead of hoping a black box behaves.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Open tools let a student with a tired laptop compete.&lt;/strong&gt; CCXT, FastAPI, React and Gemma are all free to use. That is how a student built this.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  The Lazy Payoff
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Before&lt;/strong&gt;: i spend couple of study or sleep hours a day on chart analysis.&lt;br&gt;
&lt;strong&gt;Now&lt;/strong&gt;: more time for sleep, and no violating study hours to analyze charts because now i will cover analysis reading the brief each morning.&lt;/p&gt;

&lt;p&gt;Hunter is a research and time-saving tool. It records what happened and explains it. It does not predict prices, give trading advice, or tell anyone what to buy or sell. Nothing here is financial advice.&lt;/p&gt;

&lt;h2&gt;
  
  
  What I Would Do Next
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Measure what price did after each event, using only the detected time, so predictions can be tested honestly.&lt;/li&gt;
&lt;li&gt;Add TabPFN only if it beats a simple baseline on that history.&lt;/li&gt;
&lt;li&gt;Run Gemma locally once I have hardware that can handle it.&lt;/li&gt;
&lt;li&gt;Add login so the watchlist and top five can sync across devices. It is on my roadmap and not built yet.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Prize Categories
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Best Use of Render&lt;/strong&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Best Use of Gemma&lt;/strong&gt;&lt;/li&gt;
&lt;/ul&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
      <category>opensource</category>
      <category>ai</category>
    </item>
    <item>
      <title>LedgerMind AI: Turning a Small Business Owner's Voice Notes into Records They Can Trust</title>
      <dc:creator>Classified</dc:creator>
      <pubDate>Sat, 17 Oct 2026 07:03:49 +0000</pubDate>
      <link>https://dev.to/iamclassified/ledgermind-ai-turning-a-small-business-owners-voice-notes-into-records-they-can-trust-5727</link>
      <guid>https://dev.to/iamclassified/ledgermind-ai-turning-a-small-business-owners-voice-notes-into-records-they-can-trust-5727</guid>
      <description>&lt;p&gt;&lt;em&gt;This is a submission for the &lt;a href="https://dev.to/challenges/hacktoberfest-week2-2026-10-12"&gt;Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle&lt;/a&gt;&lt;/em&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What I Built
&lt;/h2&gt;

&lt;p&gt;Most small business owners in Nigeria don't keep proper books. They keep notes: a message to themselves, a page in an exercise book, a voice memo. "Paid the plumber 40,000." "Client owes me 60,000." Those notes hold the real history of the business, but nobody can ask them questions like "how much did I take in this week?" or "who still owes me?"&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;LedgerMind AI&lt;/strong&gt; is a configurable business-record and intelligence platform that turns those everyday notes into structured records and honest reports. An owner types or speaks a note. A hosted Gemma model reads it and proposes a draft record. The owner reviews and confirms it. The backend then calculates every figure from the saved records. Only after that does the AI write a plain-language summary of how the business is doing.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Who would pay for it:&lt;/strong&gt; shop and restaurant owners, small contractors and farmers, who want to know where their money is going without learning accounting software or hiring a bookkeeper.&lt;/p&gt;

&lt;h3&gt;
  
  
  The idea in four principles
&lt;/h3&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Templates, not one-size-fits-all.&lt;/strong&gt; A restaurant, a supermarket and a contractor record different things. LedgerMind uses a template-driven record engine so each business type gets its own fields, kinds of records and reports from one shared backend.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;The AI drafts, the human confirms.&lt;/strong&gt; The model proposes records but never saves them, and it never confirms a payment. Nothing enters the books without the owner's approval.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Numbers come from code, words come from the model.&lt;/strong&gt; The backend computes every total from saved records. The model only writes the narrative and has no way to submit SQL or change data. If the AI is wrong or unavailable, the figures are still right.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Honest about what it can't claim.&lt;/strong&gt; Reports are calculated from saved records. Reconciliation differences are review signals, not proof of missing money or wrongdoing. This is not audited accounting.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Where Things Stand Today (October 17, 2026)
&lt;/h2&gt;

&lt;p&gt;I'm submitting this unfinished, on purpose, and I want to be clear about what is and isn't done.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Working and tested by me:&lt;/strong&gt;&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Registration, sign-in, revocable sessions, password and profile management. Passwords are PBKDF2-HMAC-SHA256 hashes and only token hashes are stored.&lt;/li&gt;
&lt;li&gt;Per-business ownership with membership checks enforced on the server.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Restaurants and supermarkets&lt;/strong&gt; work end to end: record entry, reports, the AI note interpreter, and the AI Business Analyst.&lt;/li&gt;
&lt;li&gt;Supplier invoices from the AI investigator are saved only after the user clicks "Confirm and save invoice", in a dedicated table separate from generic purchase records.&lt;/li&gt;
&lt;li&gt;Browser voice input through the Web Speech API, with typed notes as a fallback.&lt;/li&gt;
&lt;li&gt;Deployed on Render, with optional Sentry tracing.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;&lt;strong&gt;Not finished yet:&lt;/strong&gt;&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Other business types&lt;/strong&gt; such as contractors, pharmacies, farms and trucking/logistics only partly work. The record kinds, labels, totals and reports are still shaped around shops. Some records from other types are not yet counted, so their totals can show ₦0, and the Overview and Reports pages can disagree.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;The AI extraction makes mistakes outside the shop cases I tested.&lt;/strong&gt; For example, it can misread the direction of a payment (money received versus money paid) or save a placeholder word as a name. It does not yet ask a follow-up question when an amount is unclear.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;The hosted Gemma service sometimes returns 429, 500 and 503 errors,&lt;/strong&gt; and my retry logic is too weak to ride them out.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;The Business Analyst is slow&lt;/strong&gt; on some requests.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;"Profit" is not calculated.&lt;/strong&gt; I plan to answer profit questions as net cash movement (money in minus money out) with a clear note that it is not audited profit.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;I have a written list of 25 specific fixes for these. I chose to publish now rather than miss the deadline, and I will keep fixing them. Any commits after the October 18 deadline are listed in the repository README.&lt;/p&gt;

&lt;h2&gt;
  
  
  Demo
&lt;/h2&gt;

&lt;p&gt;Live app: &lt;strong&gt;&lt;a href="https://ledgermind-web.onrender.com" rel="noopener noreferrer"&gt;ledgermind-web.onrender.com&lt;/a&gt;&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;The app runs on Render's free tier, so if it has been idle, the first load can take up to a minute while the server wakes up. Please give it a moment.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fe73mcjabewv10ewb0hhp.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fe73mcjabewv10ewb0hhp.png" alt="Screenshot of the LedgerMind AI app" width="799" height="554"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fk3ndpsaj9g129tj4xy43.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fk3ndpsaj9g129tj4xy43.png" alt="Second screenshot of the LedgerMind AI app" width="800" height="354"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Code
&lt;/h2&gt;

&lt;p&gt;GitHub repository: &lt;a href="https://github.com/IamClassified9/LedgerMind-Ai" rel="noopener noreferrer"&gt;IamClassified9/LedgerMind-Ai&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Development began on October 13, 2026. The repository was initialized on October 16, 2026.&lt;/p&gt;

&lt;h2&gt;
  
  
  How I Built It
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Backend:&lt;/strong&gt; FastAPI, SQLAlchemy, Alembic and PostgreSQL. A shared template-driven record engine supports multiple business types.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Frontend:&lt;/strong&gt; React, TypeScript and Vite.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;AI:&lt;/strong&gt; Gemma through hosted inference, used for three jobs: interpreting typed or spoken notes into draft records, extracting supplier invoices for review, and writing the Business Analyst narrative over a snapshot the backend computes.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Guardrails:&lt;/strong&gt; The model has no database access. Its output goes through review, so a wrong answer becomes a draft a person can reject, not a wrong ledger entry.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Hosting and monitoring:&lt;/strong&gt; Render, with Sentry tracing on the AI calls.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  What Sentry Showed Me
&lt;/h2&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fzkj3s2etoa4z1lfx20dd.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fzkj3s2etoa4z1lfx20dd.png" alt="Sentry issues list showing Gemma provider 503, 429, timeout and output token limit errors on the interpret-record and analysis endpoints" width="800" height="367"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Sentry grouped the failures on my AI endpoints into separate issues. In the 14-day view, with everything first seen in the last day or so:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Gemma provider returned HTTP 503&lt;/strong&gt; on &lt;code&gt;/api/v1/ai/interpret-record&lt;/code&gt;: 7 events, the most recent about three hours before my screenshot.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Gemma provider returned HTTP 429&lt;/strong&gt; on the same endpoint: 3 events.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Gemma hit its output token limit while writing the analysis&lt;/strong&gt; on &lt;code&gt;/api/v1/ai/businesses/{business_id}/analysis&lt;/code&gt;: 1 event.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Gemma inference timed out&lt;/strong&gt; on &lt;code&gt;interpret-record&lt;/code&gt;: 1 event.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;That list taught me four things about my own code:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Most failures are the provider being busy, and they land on the path people use most.&lt;/strong&gt; Note interpretation is the first thing a user does, and it is where the 503s and 429s hit. My retries waited only 1 and then 2 seconds, and in my logs all three attempts failed within 3 seconds of each other.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;The Business Analyst's output limit was too low.&lt;/strong&gt; One analysis was cut off at the model's output token limit, which I had set to 2,500. I'm raising it to 5,000.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;One analyst request took about 45 seconds.&lt;/strong&gt; My logs showed it, and that is far too long for an owner waiting on a phone.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;The issue titles tell me the status code but not the reason.&lt;/strong&gt; Even when the logs showed Google's reason, it was a generic "Internal error." I need to record the attempt number, model and latency on each failure so it explains itself.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Next, I'll add exponential backoff with jitter, honor &lt;code&gt;Retry-After&lt;/code&gt;, make each attempt its own span, and send the model a smaller snapshot.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why Does Open Innovation Matter?
&lt;/h2&gt;

&lt;p&gt;I used Gemma through hosted inference, and I want to be upfront that this is a trade-off. Nobody has to own a GPU to use the app, but a customer's note does leave the app to reach the provider. LedgerMind AI does not claim that data stays on infrastructure I control when hosted inference is used.&lt;/p&gt;

&lt;p&gt;What the open-weight choice gives me is a way out. When the hosted endpoint returned errors that survived my retries, the fix I can reach for is to configure another provider, or self-host once a business can afford it, rather than wait for one vendor. I can also keep the model on a short leash, since it never writes SQL or saves records and the backend computes every figure. If a better or cheaper model appears, I can replace the narrator without touching the numbers.&lt;/p&gt;

&lt;p&gt;The next step where open weights matter most is fine-tuning for how these owners actually write: informal notes, currency amounts, and phrases like "client owes me" where the direction of money changes everything. A general-purpose API handles that poorly, and my own extraction mistakes show it.&lt;/p&gt;

&lt;h2&gt;
  
  
  How I Used Render
&lt;/h2&gt;

&lt;p&gt;Render hosts the deployed LedgerMind AI application, so a business owner can use it from a browser without installing the backend or running a model on their own computer.&lt;/p&gt;

&lt;p&gt;The deployment connects the React, TypeScript and Vite frontend to the FastAPI backend. When a user submits a business note, the backend handles the request and calls the hosted Gemma inference service. The model proposes a structured draft, which the user reviews before confirming it. The backend then saves the approved record and calculates the figures used in reports.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Deployment architecture&lt;/strong&gt;&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Frontend:&lt;/strong&gt; Render, at &lt;a href="https://ledgermind-web.onrender.com" rel="noopener noreferrer"&gt;ledgermind-web.onrender.com&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Backend:&lt;/strong&gt; Render, at &lt;a href="https://ledgermind-api-4guj.onrender.com" rel="noopener noreferrer"&gt;ledgermind-api-4guj.onrender.com&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Database:&lt;/strong&gt; PostgreSQL on Render&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;AI inference:&lt;/strong&gt; Gemma through Google's hosted inference&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Monitoring:&lt;/strong&gt; Sentry traces the AI calls and helps me find latency and provider errors&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Render makes the application accessible, while the external inference service supplies the model capabilities. I do not claim that Gemma itself runs on Render.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Evidence from the deployed application&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fw42uh31p434wc10ob2r1.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fw42uh31p434wc10ob2r1.png" alt="Screenshot of the Render dashboard showing the LedgerMind AI services (1 of 3)" width="800" height="528"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fjbb8gw2gql5q9y2u62fe.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fjbb8gw2gql5q9y2u62fe.png" alt="Screenshot of the Render dashboard showing the LedgerMind AI services (2 of 3)" width="799" height="371"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fn1h4nbkl7ruhfx2z7229.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fn1h4nbkl7ruhfx2z7229.png" alt="Screenshot of the Render dashboard showing the LedgerMind AI services (3 of 3)" width="800" height="715"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What I learned from Sentry&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;Deploying the app exposed a real operational issue: the hosted Gemma service sometimes returns errors or takes too long to respond, and on Render's free tier the server itself can be asleep when the first request arrives. Sentry helped me investigate the Gemma failures. My next improvements are stronger retry handling, clearer error reporting, better response times, and a friendlier "server is waking up" message.&lt;/p&gt;

&lt;p&gt;This deployment shows how a small-business AI application can be made available without requiring its users to operate their own model-serving infrastructure.&lt;/p&gt;

&lt;h2&gt;
  
  
  Prize Categories
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Best Use of Gemma:&lt;/strong&gt; Hosted Gemma interprets notes, extracts invoices and writes the Business Analyst narrative.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Best Use of Render:&lt;/strong&gt; The frontend and backend are deployed on Render, connected to a hosted Gemma service. See "How I Used Render" above.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Best Use of Sentry Agent Tracing:&lt;/strong&gt; Sentry traces the Gemma calls, which showed me my weak retries, a 45-second analyst request and vague error reporting.&lt;/li&gt;
&lt;/ul&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
      <category>ai</category>
      <category>opensource</category>
    </item>
    <item>
      <title>Gemma</title>
      <dc:creator>Classified</dc:creator>
      <pubDate>Thu, 15 Oct 2026 11:53:59 +0000</pubDate>
      <link>https://dev.to/iamclassified/gemma-14k2</link>
      <guid>https://dev.to/iamclassified/gemma-14k2</guid>
      <description>&lt;p&gt;Guys im trying to use Gemma and it keeps giving me internal error 500, what to do about this? is Gemma down or what ?&lt;/p&gt;

</description>
      <category>ai</category>
      <category>opensource</category>
      <category>gemma</category>
    </item>
    <item>
      <title>I Published CLASSIFIED: A Python Package for Shared Market-Data Infrastructure</title>
      <dc:creator>Classified</dc:creator>
      <pubDate>Mon, 12 Oct 2026 02:40:19 +0000</pubDate>
      <link>https://dev.to/iamclassified/i-published-classified-a-python-package-for-shared-market-data-infrastructure-5h9j</link>
      <guid>https://dev.to/iamclassified/i-published-classified-a-python-package-for-shared-market-data-infrastructure-5h9j</guid>
      <description>&lt;p&gt;I’ve published &lt;strong&gt;CLASSIFIED&lt;/strong&gt; on PyPI under the distribution name &lt;code&gt;classifiedai&lt;/code&gt;.&lt;/p&gt;

&lt;p&gt;The idea behind the project is straightforward: applications that work with market data shouldn't all need to implement their own exchange connectivity and data-handling infrastructure.&lt;/p&gt;

&lt;p&gt;CLASSIFIED provides a shared foundation for retrieving and handling market data, with support for normalized OHLCV data, multi-timeframe requests, and HTTP/WebSocket delivery through its API gateway.&lt;/p&gt;

&lt;h3&gt;
  
  
  What it provides
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;Exchange connectivity through CCXT.&lt;/li&gt;
&lt;li&gt;Normalized OHLCV candle data.&lt;/li&gt;
&lt;li&gt;Multi-timeframe market-data requests.&lt;/li&gt;
&lt;li&gt;Historical market-data retrieval.&lt;/li&gt;
&lt;li&gt;HTTP endpoints and WebSocket streaming through the gateway.&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  Installation
&lt;/h3&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;python &lt;span class="nt"&gt;-m&lt;/span&gt; pip &lt;span class="nb"&gt;install &lt;/span&gt;classifiedai
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The PyPI distribution is named &lt;code&gt;classifiedai&lt;/code&gt;, but the Python import namespace remains &lt;code&gt;classified&lt;/code&gt;.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight python"&gt;&lt;code&gt;&lt;span class="kn"&gt;import&lt;/span&gt; &lt;span class="n"&gt;classified&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h3&gt;
  
  
  Links
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;PyPI package:&lt;/strong&gt; &lt;a href="https://pypi.org/project/classifiedai/" rel="noopener noreferrer"&gt;https://pypi.org/project/classifiedai/&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;GitHub repository and README:&lt;/strong&gt; &lt;a href="https://github.com/IamClassified9/Project-Documentation/blob/main/CLASSIFIED" rel="noopener noreferrer"&gt;https://github.com/IamClassified9/Project-Documentation/blob/main/CLASSIFIED&lt;/a&gt;
&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  Why I built it
&lt;/h3&gt;

&lt;p&gt;I wanted a reusable market-data foundation that applications and analytical systems could consume without each one having to manage its own exchange integration.&lt;/p&gt;

&lt;p&gt;CLASSIFIED is infrastructure, not a trading bot or a trading strategy. Its purpose is to provide market-data functionality that other systems can build on.&lt;/p&gt;

&lt;p&gt;The package is now published, and I’m interested in feedback from Python developers working with exchange APIs, OHLCV data, charting, and real-time market-data systems.&lt;/p&gt;

&lt;p&gt;If you check it out, I’d appreciate your thoughts on the package structure, API design, and areas worth improving.&lt;/p&gt;

</description>
      <category>python</category>
      <category>opensource</category>
      <category>cryptocurrency</category>
      <category>programming</category>
    </item>
    <item>
      <title>My Intro</title>
      <dc:creator>Classified</dc:creator>
      <pubDate>Mon, 12 Oct 2026 02:35:17 +0000</pubDate>
      <link>https://dev.to/iamclassified/my-intro-4ean</link>
      <guid>https://dev.to/iamclassified/my-intro-4ean</guid>
      <description>&lt;h1&gt;
  
  
  👋 Hello DEV Community,  I'm Mathew, and I'm building ClassifiedLens
&lt;/h1&gt;

&lt;p&gt;Hey everyone! 👋&lt;/p&gt;

&lt;p&gt;I'm Mathew, a developer interested in building practical software products, developer tools, backend systems, APIs, automation, and modern web applications.&lt;/p&gt;

&lt;p&gt;I'm excited to finally join the DEV Community and start sharing what I'm building, what I'm learning, and the problems I'm trying to solve.&lt;/p&gt;

&lt;h2&gt;
  
  
  🚀 One of the things I'm currently building: ClassifiedLens
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;ClassifiedLens&lt;/strong&gt; is a TypeScript/React visualization SDK I'm building for developers who need to integrate rich financial and market visualizations into their applications.&lt;/p&gt;

&lt;p&gt;The idea is simple:&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;**Developers should be able to tell the visualization system what they want to display without having to build the entire rendering system themselves.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;Instead of rebuilding charting and visualization infrastructure from scratch, ClassifiedLens is designed to provide reusable visualization capabilities that applications can integrate into their own systems.&lt;/p&gt;

&lt;h3&gt;
  
  
  🧩 What I'm working on
&lt;/h3&gt;

&lt;p&gt;Some of the areas I'm exploring include:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;📊 Market and financial data visualization&lt;/li&gt;
&lt;li&gt;📈 Interactive charts&lt;/li&gt;
&lt;li&gt;🎯 Custom overlays&lt;/li&gt;
&lt;li&gt;🔌 Developer-friendly APIs&lt;/li&gt;
&lt;li&gt;⚛️ React integration&lt;/li&gt;
&lt;li&gt;📦 TypeScript/JavaScript SDK distribution&lt;/li&gt;
&lt;li&gt;🧱 Reusable visualization components&lt;/li&gt;
&lt;li&gt;🔄 Real-time data visualization&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The project is still evolving, and that's actually one of the reasons I'm excited to share it here.&lt;/p&gt;

&lt;p&gt;I want to build it in public, learn from other developers, and improve the project based on real-world use cases and feedback.&lt;/p&gt;

&lt;h2&gt;
  
  
  📦 It's already available as an npm package
&lt;/h2&gt;

&lt;p&gt;I've published the first version of &lt;strong&gt;ClassifiedLens&lt;/strong&gt; to npm.&lt;/p&gt;

&lt;p&gt;I'm currently working on improving the SDK, documentation, developer experience, and the overall architecture.&lt;/p&gt;

&lt;p&gt;If you're interested in financial applications, trading interfaces, data visualization, React, TypeScript, or developer tools, I'd love to hear your thoughts.&lt;/p&gt;

&lt;h2&gt;
  
  
  👨‍💻 What I'm hoping to do here
&lt;/h2&gt;

&lt;p&gt;This isn't just about promoting one project.&lt;/p&gt;

&lt;p&gt;I want to use DEV to:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Share what I'm learning&lt;/li&gt;
&lt;li&gt;Document my development process&lt;/li&gt;
&lt;li&gt;Talk about architecture and engineering decisions&lt;/li&gt;
&lt;li&gt;Showcase projects I'm building&lt;/li&gt;
&lt;li&gt;Learn from other developers&lt;/li&gt;
&lt;li&gt;Get feedback from people who have solved similar problems&lt;/li&gt;
&lt;li&gt;Connect with developers and builders around the world&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;I'm particularly interested in meeting other developers building SaaS products, developer tools, fintech applications, APIs, automation systems, and open-source projects.&lt;/p&gt;

&lt;p&gt;So...&lt;/p&gt;

&lt;p&gt;If you're interested in what I'm building, check out &lt;strong&gt;ClassifiedLens&lt;/strong&gt; and let me know what you think.&lt;/p&gt;

&lt;p&gt;🚀 Let's build.&lt;/p&gt;

&lt;h1&gt;
  
  
  javascript #typescript #react #webdev #opensource #devtools #datavisualization #fintech
&lt;/h1&gt;

</description>
      <category>ai</category>
      <category>webdev</category>
      <category>programming</category>
      <category>cryptocurrency</category>
    </item>
    <item>
      <title>[Boost]</title>
      <dc:creator>Classified</dc:creator>
      <pubDate>Fri, 09 Oct 2026 23:08:29 +0000</pubDate>
      <link>https://dev.to/iamclassified/-1315</link>
      <guid>https://dev.to/iamclassified/-1315</guid>
      <description>&lt;div class="ltag__link--embedded"&gt;
  &lt;div class="crayons-story "&gt;
  &lt;a href="https://dev.to/iamclassified/liquidity-event-identificationan-llm-reasoning-benchmark-5boi" class="crayons-story__hidden-navigation-link"&gt;LIQUIDITY EVENT IDENTIFICATION An LLM Reasoning Benchmark&lt;/a&gt;


  &lt;div class="crayons-story__body crayons-story__body-full_post"&gt;
      &lt;a href="https://dev.to/iamclassified/liquidity-event-identificationan-llm-reasoning-benchmark-5boi" class="crayons-article__context-note crayons-article__context-note__feed"&gt;&lt;p&gt;Kaggle Benchmarking Challenge Submission&lt;/p&gt;

&lt;/a&gt;
    &lt;div class="crayons-story__top"&gt;
      &lt;div class="crayons-story__meta"&gt;
        &lt;div class="crayons-story__author-pic"&gt;

          &lt;a href="/iamclassified" class="crayons-avatar  crayons-avatar--l  "&gt;
            &lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4143571%2F87426864-f60a-4480-b3e5-a0b8c21aab5c.jpg" alt="iamclassified profile" class="crayons-avatar__image"&gt;
          &lt;/a&gt;
        &lt;/div&gt;
        &lt;div&gt;
          &lt;div&gt;
            &lt;a href="/iamclassified" class="crayons-story__secondary fw-medium m:hidden"&gt;
              Classified
            &lt;/a&gt;
            &lt;div class="profile-preview-card relative mb-4 s:mb-0 fw-medium hidden m:inline-block"&gt;
              
                Classified
                
                
              
              &lt;div id="story-author-preview-content-4793629" class="profile-preview-card__content crayons-dropdown branded-7 p-4 pt-0"&gt;
                &lt;div class="gap-4 grid"&gt;
                  &lt;div class="-mt-4"&gt;
                    &lt;a href="/iamclassified" class="flex"&gt;
                      &lt;span class="crayons-avatar crayons-avatar--xl mr-2 shrink-0"&gt;
                        &lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4143571%2F87426864-f60a-4480-b3e5-a0b8c21aab5c.jpg" class="crayons-avatar__image" alt=""&gt;
                      &lt;/span&gt;
                      &lt;span class="crayons-link crayons-subtitle-2 mt-5"&gt;Classified&lt;/span&gt;
                    &lt;/a&gt;
                  &lt;/div&gt;
                  &lt;div class="print-hidden"&gt;
                    
                      Follow
                    
                  &lt;/div&gt;
                  &lt;div class="author-preview-metadata-container"&gt;&lt;/div&gt;
                &lt;/div&gt;
              &lt;/div&gt;
            &lt;/div&gt;

          &lt;/div&gt;
          &lt;a href="https://dev.to/iamclassified/liquidity-event-identificationan-llm-reasoning-benchmark-5boi" class="crayons-story__tertiary fs-xs"&gt;&lt;time&gt;Oct 3&lt;/time&gt;&lt;span class="time-ago-indicator-initial-placeholder"&gt;&lt;/span&gt;&lt;/a&gt;
        &lt;/div&gt;
      &lt;/div&gt;

    &lt;/div&gt;

    &lt;div class="crayons-story__indention"&gt;
      &lt;h2 class="crayons-story__title crayons-story__title-full_post"&gt;
        &lt;a href="https://dev.to/iamclassified/liquidity-event-identificationan-llm-reasoning-benchmark-5boi" id="article-link-4793629"&gt;
          LIQUIDITY EVENT IDENTIFICATION An LLM Reasoning Benchmark
        &lt;/a&gt;
      &lt;/h2&gt;
        &lt;div class="crayons-story__tags"&gt;
            &lt;a class="crayons-tag  crayons-tag--monochrome " href="/t/devchallenge"&gt;&lt;span class="crayons-tag__prefix"&gt;#&lt;/span&gt;devchallenge&lt;/a&gt;
            &lt;a class="crayons-tag  crayons-tag--monochrome " href="/t/kagglechallenge"&gt;&lt;span class="crayons-tag__prefix"&gt;#&lt;/span&gt;kagglechallenge&lt;/a&gt;
            &lt;a class="crayons-tag  crayons-tag--monochrome " href="/t/ai"&gt;&lt;span class="crayons-tag__prefix"&gt;#&lt;/span&gt;ai&lt;/a&gt;
            &lt;a class="crayons-tag  crayons-tag--monochrome " href="/t/machinelearning"&gt;&lt;span class="crayons-tag__prefix"&gt;#&lt;/span&gt;machinelearning&lt;/a&gt;
        &lt;/div&gt;
      &lt;div class="crayons-story__bottom"&gt;
        &lt;div class="crayons-story__details"&gt;
          &lt;a href="https://dev.to/iamclassified/liquidity-event-identificationan-llm-reasoning-benchmark-5boi" class="crayons-btn crayons-btn--s crayons-btn--ghost crayons-btn--icon-left"&gt;
            &lt;div class="multiple_reactions_aggregate"&gt;
              &lt;span class="multiple_reactions_icons_container"&gt;
                  &lt;span class="crayons_icon_container"&gt;
                    &lt;img src="https://assets.dev.to/assets/raised-hands-74b2099fd66a39f2d7eed9305ee0f4553df0eb7b4f11b01b6b1b499973048fe5.svg" width="18" height="18"&gt;
                  &lt;/span&gt;
                  &lt;span class="crayons_icon_container"&gt;
                    &lt;img src="https://assets.dev.to/assets/multi-unicorn-b44d6f8c23cdd00964192bedc38af3e82463978aa611b4365bd33a0f1f4f3e97.svg" width="18" height="18"&gt;
                  &lt;/span&gt;
                  &lt;span class="crayons_icon_container"&gt;
                    &lt;img src="https://assets.dev.to/assets/sparkle-heart-5f9bee3767e18deb1bb725290cb151c25234768a0e9a2bd39370c382d02920cf.svg" width="18" height="18"&gt;
                  &lt;/span&gt;
              &lt;/span&gt;
              &lt;span class="aggregate_reactions_counter"&gt;3&lt;span class="hidden s:inline"&gt;&amp;nbsp;reactions&lt;/span&gt;&lt;/span&gt;
            &lt;/div&gt;
          &lt;/a&gt;
            &lt;a href="https://dev.to/iamclassified/liquidity-event-identificationan-llm-reasoning-benchmark-5boi#comments" class="crayons-btn crayons-btn--s crayons-btn--ghost crayons-btn--icon-left flex items-center"&gt;
              

              &lt;span class="hidden s:inline"&gt;Add&amp;nbsp;Comment&lt;/span&gt;
            &lt;/a&gt;
        &lt;/div&gt;
        &lt;div class="crayons-story__save"&gt;
          &lt;small class="crayons-story__tertiary fs-xs mr-2"&gt;
            2 min read
          &lt;/small&gt;
        &lt;/div&gt;
      &lt;/div&gt;
    &lt;/div&gt;
  &lt;/div&gt;
&lt;/div&gt;

&lt;/div&gt;


</description>
      <category>ai</category>
      <category>llm</category>
      <category>machinelearning</category>
    </item>
    <item>
      <title>LIQUIDITY EVENT IDENTIFICATION An LLM Reasoning Benchmark</title>
      <dc:creator>Classified</dc:creator>
      <pubDate>Sat, 03 Oct 2026 21:45:44 +0000</pubDate>
      <link>https://dev.to/iamclassified/liquidity-event-identificationan-llm-reasoning-benchmark-5boi</link>
      <guid>https://dev.to/iamclassified/liquidity-event-identificationan-llm-reasoning-benchmark-5boi</guid>
      <description>&lt;p&gt;&lt;em&gt;This is a submission for the &lt;a href="https://dev.to/challenges/kaggle-2026-09-23"&gt;Kaggle Benchmarking Challenge&lt;/a&gt;&lt;/em&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What I Benchmarked
&lt;/h2&gt;

&lt;p&gt;I built a Kaggle benchmark called &lt;strong&gt;Liquidity Event Identification&lt;/strong&gt; to test how well large language models can identify liquidity events from controlled price-action scenarios.&lt;/p&gt;

&lt;p&gt;The benchmark tests whether a model can distinguish between events such as liquidity sweeps, rejections, breakouts, and acceptance based on the evidence provided in each scenario.&lt;/p&gt;

&lt;p&gt;The benchmark contains seven scenarios:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Equal-high liquidity sweep and rejection&lt;/li&gt;
&lt;li&gt;Equal-low liquidity sweep and rejection&lt;/li&gt;
&lt;li&gt;Breakout followed by acceptance&lt;/li&gt;
&lt;li&gt;Brief break followed by rejection&lt;/li&gt;
&lt;li&gt;Near-identical scenarios where one detail changes the correct interpretation&lt;/li&gt;
&lt;li&gt;Conflicting multi-timeframe evidence&lt;/li&gt;
&lt;li&gt;A deliberately misleading explanation&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The benchmark is focused on reasoning about market structure and liquidity events. It is not designed to measure trading profitability or predict real market outcomes.&lt;/p&gt;

&lt;h2&gt;
  
  
  Models Tested
&lt;/h2&gt;

&lt;p&gt;I ran the benchmark against 20 models available through Kaggle:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;GPT-6 Astra&lt;/li&gt;
&lt;li&gt;GPT-5.6 Sol&lt;/li&gt;
&lt;li&gt;GPT-5.6 Luna&lt;/li&gt;
&lt;li&gt;GPT-5.5&lt;/li&gt;
&lt;li&gt;GPT-5.4&lt;/li&gt;
&lt;li&gt;GPT-5.4 Mini&lt;/li&gt;
&lt;li&gt;GPT-5.4 Nano&lt;/li&gt;
&lt;li&gt;Claude Opus 5&lt;/li&gt;
&lt;li&gt;Claude Opus 4.6&lt;/li&gt;
&lt;li&gt;Claude Sonnet 5&lt;/li&gt;
&lt;li&gt;Gemini 3.7 Flash&lt;/li&gt;
&lt;li&gt;Gemini 3 Flash Preview&lt;/li&gt;
&lt;li&gt;Gemini 3.1 Flash-Lite Preview&lt;/li&gt;
&lt;li&gt;Gemini 3.1 Pro Preview&lt;/li&gt;
&lt;li&gt;Gemini 3.5 Flash&lt;/li&gt;
&lt;li&gt;Grok 4.20&lt;/li&gt;
&lt;li&gt;Qwen3 Next 80B&lt;/li&gt;
&lt;li&gt;GPT-OSS 120B&lt;/li&gt;
&lt;li&gt;DeepSeek R1&lt;/li&gt;
&lt;li&gt;Claude Sonnet 4.6&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;I used models from multiple providers to compare how different models performed on the same scenarios.&lt;/p&gt;

&lt;h2&gt;
  
  
  Findings
&lt;/h2&gt;

&lt;p&gt;16 of the 20 model runs completed successfully. All 16 completed runs scored &lt;strong&gt;7/7&lt;/strong&gt;, resulting in &lt;strong&gt;112 correct answers out of 112 completed scenario evaluations&lt;/strong&gt;.&lt;/p&gt;

&lt;p&gt;The other four runs — Qwen3 Next 80B, GPT-OSS 120B, DeepSeek R1, and Claude Sonnet 4.6 — encountered execution errors before completing the benchmark. They were excluded from the accuracy calculation.&lt;/p&gt;

&lt;p&gt;The main result was that the benchmark did not differentiate the completed models. Every completed model achieved a perfect score.&lt;/p&gt;

&lt;p&gt;This suggests that the seven scenarios are not difficult enough to distinguish model performance.&lt;/p&gt;

&lt;p&gt;The 100% result also does not indicate that these models can trade profitably or reliably interpret real-world market conditions. The scenarios are controlled and synthetic, and the benchmark measures a specific reasoning task.&lt;/p&gt;

&lt;p&gt;For the next version, I plan to increase the difficulty with more scenarios involving ambiguity, conflicting evidence, misleading context, stronger multi-timeframe conflicts, and cases where the most obvious interpretation is incorrect.&lt;/p&gt;

&lt;h2&gt;
  
  
  My Benchmark
&lt;/h2&gt;

&lt;p&gt;My Kaggle benchmark:&lt;/p&gt;

&lt;p&gt;&lt;a href="https://www.kaggle.com/benchmarks/tasks/iamclassified/liquidity-event-identification" rel="noopener noreferrer"&gt;Liquidity Event Identification — Kaggle&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;The benchmark contains the task definition, scenarios, evaluation logic, and model evaluation results.&lt;/p&gt;

</description>
      <category>devchallenge</category>
      <category>kagglechallenge</category>
      <category>ai</category>
      <category>machinelearning</category>
    </item>
  </channel>
</rss>
