The latest articles on DEV Community by Arfath Pasha (@arfath_pasha_). https://dev.to/arfath_pasha_ 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%2F3762426%2Fcdc52c5b-ed31-475d-a3a2-42d631f13784.jpg https://dev.to/arfath_pasha_ en Arfath Pasha Sat, 11 Jul 2026 10:01:33 +0000 https://dev.to/arfath_pasha_/obsession-i-built-an-obsession-tracker-for-developers-5g60 https://dev.to/arfath_pasha_/obsession-i-built-an-obsession-tracker-for-developers-5g60 <p><em>This is a submission for <a href="https://dev.to/challenges/weekend-2026-07-09">Weekend Challenge: Passion Edition</a></em></p> <h2> What I Built </h2> <p>I built <strong>TOTAL_OBSESSION_DETECTOR (STRAT_OS)</strong>, a Neo-Brutalist "Editorial Aesthetic" passion &amp; neurosis analyzer designed specifically for developer subcultures. </p> <p>Whether it's spending $600 tuning custom mechanical keyboard switches, building multi-node Kubernetes home labs in a closet, or obsessively optimizing database index selectivities, developers have high-severity fixations. <strong>TOTAL_OBSESSION_DETECTOR</strong> lets you log these behavioral symtoms, run dynamic, AI-powered "Passion Audits" to diagnose your obsession level (complete with a brutal roast and an intellectual validation manifesto), and instantly generate a <strong>Quick Defense Shield</strong>—a hyper-formal, jargon-rich legalistic defense letter to dismiss questioning from supervisors, landlords, or partners.</p> <h2> Demo </h2> <p>🔗 <strong>Live Deployed App</strong>: <a href="https://obsession-tracker.ai.studio" rel="noopener noreferrer">TOTAL_OBSESSION_DETECTOR</a> </p> <h2> Code </h2> <p>🔗 <strong>GitHub Repository</strong>: <em>[<a href="https://github.com/arfath62/obsession-" rel="noopener noreferrer">https://github.com/arfath62/obsession-</a>]</em></p> <h2> How I Built It </h2> <p>I engineered a full-stack Node.js &amp; React single-page application focused on high performance and extreme resilience:</p> <ol> <li> <strong>Editorial Design System</strong>: Kept the layout true to print design and neo-brutalist aesthetics—featuring hard geometric lines, deep blacks, spacious negative layouts, raw code readouts, and responsive transitions using <code>motion</code>.</li> <li> <strong>Resilient AI Pipeline</strong>: Implemented the official Google GenAI SDK (<code>@google/genai</code>) inside an Express.js backend. To safeguard the application from high-demand <code>503 UNAVAILABLE</code> errors or rate-limiting: <ul> <li>Configured an active retry protocol using exponential backoffs.</li> <li>Built an automated model-fallback waterfall. If <code>gemini-3.5-flash</code> experiences transient load, the server seamlessly cascades down to <code>gemini-flash-latest</code> and then <code>gemini-3.1-flash-lite</code>, assuring 100% backend uptime.</li> </ul> </li> <li> <strong>Structured Outputs</strong>: Forced Gemini to spit out structured, schema-validated JSON diagnostics so the client could instantly render progress gauges, witty deconstruction blocks, and formatted procedural steps.</li> </ol> <h2> Prize Categories </h2> <h3> Best Use of Google AI </h3> <p>I integrated the official <code>@google/genai</code> SDK and established a multi-model fallback waterfall system (<code>gemini-3.5-flash</code> ➡️ <code>gemini-flash-latest</code> ➡️ <code>gemini-3.1-flash-lite</code>) to bypass high-demand unavailability spikes, assuring that developer audits and legal defense generation are always available.<br> thank you....</p> ai devchallenge webdev programming