DEV Community: Toluwaloju Kayode

Localbiz Growth. A content creator you can trust

Toluwaloju Kayode — Mon, 19 Oct 2026 02:35:29 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle

What I Built

This is a Online local business content creation tool. It helps local businesses to create online content where they can post videos and images of what they do and application creates a top viewed content and posts it to any of their connected social media account. It helps local businesses no longer worry about what to post as the application handles all the posting details.
Local businesses pay for the application and its use on their andriod and ios devices.

Demo

expo.dev

The above link is the Android application of the localbiz application built on npx expo and fastapi python backend

Code

Toluwaloju0 / business-scaler

This repo is used to house the codes for a business social presence grower

How I Built It

Why Does Open Innovation Matter?

Innovation matters for this project as the businesses do not need any subscription or api key to use the basic services provided making it possible for small businesses to reach a wide range of clients using top notch and free AI models while getting the same quality as other paid AI clients

My Agent Session

Building and Serving a LocalBiz Android APK

You

Prepare the LocalBiz Expo app for a downloadable Android APK served by the FastAPI backend, using http://[REDACTED] as the backend base URL.

Agent

Checked Expo SDK 57, Node 22, dependency compatibility, EAS build configuration, backend routes, and Android build prerequisites. Configured an EAS internal APK profile with the backend URL, a stable Android package ID, and Android cleartext support for the current HTTP backend. Added a backend-served /download page and an APK file endpoint, corrected duplicate FastAPI app construction so startup configuration validation remains active, and tested both the backend routes and Android JS bundle.

Agent

Backend tests, TypeScript, Expo dependency checks, and Android bundle export passed. The live backend serves /download and /health. The APK itself was not generated because EAS requires an Expo account login and no authenticated account is available in this environment; the host also lacks local Android SDK/ADB/Gradle. The download page reports the artifact as unavailable until a completed APK is placed in the configured build directory.

My First User

My first user provided me with the following screenshot of him using the application on EXPO GO

Prize Categories

Best Use of GitHub Copilot

Schedule Reminder. Remember your plans and leave your screen to do it.

Toluwaloju Kayode — Sun, 11 Oct 2026 22:07:04 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

What I Built

I built a voice-assisted schedule reminder application that helps users stay organized and follow through on their planned activities. It keeps track of scheduled tasks and reminds users when it is time to act by playing voice reminders. The application also listens for verbal confirmation to determine whether the user has started the scheduled activity.

The goal is to help users spend less time glued to their screens and more time taking action in the real world. By combining voice reminders with audio-based confirmation, the app encourages users to stay on track, manage their time effectively, and focus on the tasks they planned ahead of time.

Demo

This is a link where you can get the mobile application from.

mlh-touch-grass.onrender.com

NOTE: the application works on mobile devices only
and all LLM downloads occur upon first installation and opening of the application so ensure you have a minimum of 500MB of data

Code

Toluwaloju0 / mlh_touch_grass

This repository is for the mlh project touch grass. i have tried creating an executable which users can run on their local machine or devices

How I Built It

I built the application using Python 3.12 in a virtual environment, with the Codex coding agent assisting me throughout the development process.

For the voice features, I integrated Kokoro ONNX for text-to-speech, allowing the application to deliver scheduled reminders through voice playback. I also used FLAN-T5-Small as part of the language-processing pipeline for handling user confirmations.

I developed the mobile application's user interface with Kivy, a Python framework for building cross-platform applications, and used Buildozer to package the project for Android.

My goal was to create a lightweight, voice-assisted application that helps users manage their schedules, respond to reminders, and follow through on their planned activities without constantly interacting with their screens.

Why Does Open Innovation Matter?

Open innovation makes it possible to bring AI-powered voice reminders to Android devices without relying on constant internet access or expensive cloud APIs. My project aims to help users stay on schedule, spend less time on screens, and take action on their planned activities while making AI more accessible to people with limited connectivity.

My Agent Session

Voice Schedule: Offline Kivy Reminder App

You

I want to build a mobile app using Python and Kivy. Create a voice reminder app based on schedules saved by the user. Support one schedule for all days or individual schedules per weekday via a toggle. Store schedules in JSON. Include Home navigation, a schedule editor, and a Usage page. At schedule time, speak the activity and listen for one minute for agreement. Use Kokoro ONNX for speech output and an offline speech-to-text model; keep processing offline after first-run model downloads.

Agent

Implemented a Kivy app in main.py with Home, Schedule, and Usage pages; daily or weekday-specific schedules saved to local JSON; first-run downloads for Kokoro ONNX and Vosk; and local yes/no phrase classification. FLAN-T5-small is not an ASR model, so Vosk was chosen for speech recognition. The in-process checker is only best-effort: reliable Android background reminders need a foreground service, alarm and reboot handling, permissions, and Android-compatible audio/ONNX dependencies.

You

Edit buildozer.spec so this application can be packaged as a mobile .apk.

Agent

Updated buildozer.spec for the Voice Schedule app, Android INTERNET and RECORD_AUDIO permissions, API 35/minimum API 24, ARM architectures, source exclusions, and APK outputs for debug and release. Kokoro ONNX, Vosk, and sounddevice were not added as requirements because their Android compiled dependencies need compatible builds or python-for-android recipes. A debug build can be started with: buildozer -v android debug. No APK build was run.

You

How do I create a DevRelay agent session for this chat?

Agent

Explained that the user can ask Codex to save the chat as a DevRelay agent session, or submit a normalized transcript using the DevRelay CLI.

You

Save this chat as a DevRelay agent session. Follow this DEV submission format: What I Built (what it does, how it gets people off-screen and outdoors, who it is for); Demo; Code; How I Built It (open-source AI and local inference); Why Open Innovation Matters; My Agent Session; Prize Categories. Note: no demo link or repository was provided.

Schedule Reminder. Remember your plans and leave your screen to do it.

Toluwaloju Kayode — Sun, 11 Oct 2026 22:05:23 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

What I Built

Demo

This is a link where you can get the mobile application from.

mlh-touch-grass.onrender.com

NOTE: the application works on mobile devices only
and all LLM downloads occur upon first installation and opening of the application so ensure you have a minimum of 500MB of data

Code

Toluwaloju0 / mlh_touch_grass

This repository is for the mlh project touch grass. i have tried creating an executable which users can run on their local machine or devices

How I Built It

I built the application using Python 3.12 in a virtual environment, with the Codex coding agent assisting me throughout the development process.

I developed the mobile application's user interface with Kivy, a Python framework for building cross-platform applications, and used Buildozer to package the project for Android.

Why Does Open Innovation Matter?

My Agent Session

Voice Schedule: Offline Kivy Reminder App

You

Agent

You

Edit buildozer.spec so this application can be packaged as a mobile .apk.

Agent

You

How do I create a DevRelay agent session for this chat?

Agent

Explained that the user can ask Codex to save the chat as a DevRelay agent session, or submit a normalized transcript using the DevRelay CLI.

You

My sister needs help cooking meals in the kitchen. Agents which read out recipes and provide feedback and improvements should assist her

Toluwaloju Kayode — Sun, 04 Oct 2026 21:24:05 +0000

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend

What I Built
I built Mise en Place, an AI-powered culinary sous-chef and interactive recipe workbench designed for a close friend who recently moved into their first apartment and wanted to stop relying on expensive takeout and start cooking real meals at home—especially cherished family recipes and authentic regional dishes they find online.

Whenever my friend tried to cook, three real-world kitchen frustrations consistently got in their way:

Scattered, Unusable Recipe Formats: Family recipes were trapped in photos of handwritten index cards or messy text messages, while online recipes were buried under 15 paragraphs of blog backstory, pop-ups, and ads.
Kitchen Math & Scaling Anxiety: Cooking for one or two people when a recipe is written for six meant doing fraction arithmetic (1/3 of 1 ¾ cups) and converting US volume cups into metric grams while food was actively sizzling on the stove.
Mid-Cook Panic with Messy Hands: Halfway through cooking, they would realize they were missing a key ingredient (like mirin or dry white wine) or didn't know what a sensory cue like "cook until translucent" or "al dente all'onda" actually looked like—and with flour- or oil-covered hands, scrolling on a phone screen was a nightmare.
Mise en Place solves all three problems by turning any recipe into a calm, hands-free kitchen station:

Multi-Format Recipe Upload (+ Upload Recipe): Upload a photo/screenshot of a cookbook page or handwritten recipe card, select a .txt/.md/.json/.pdf recipe file, or paste raw kitchen notes. The agent extracts categorized ingredients, exact numeric quantities, temperatures, and step-by-step instructions with embedded timer durations.
Live Web Recipe Sourcing & Python Schema.org Scraper (Web Sourcing): Search the web by dish name or by the ingredients currently sitting in your fridge, or paste any recipe URL. Our Python agent (scrape_recipe_url) parses Schema.org Recipe graphs directly from webpage HTML to strip away blog clutter while preserving clickable source citations. Deterministic Python Fraction & Density Scaling (Cook Mode): Scale any recipe from 1 to 24 servings and toggle between Original, Metric (g / ml / °C), and Imperial (oz / cup / °F) with real-time oven temperature conversions inside the step text. Smart Multi-Timer Dock & Hands-Free Voice Read-Aloud: Every time-bound cooking step generates a one-click countdown timer with Web Audio API chimes so multiple stove and oven stages can run simultaneously, plus a Read Step Aloud voice button for hands-free listening when your hands are covered in dough. Contextual Gemma 3 Sous-Chef Sidebar: Ask real-time culinary questions ("What can I substitute for white miso?", "What sensory cues tell me this step is done?", "How can I prep this dish ahead of time?") with full awareness of your active recipe, current step number, and scaled serving size. Demo <a href="https://www.loom.com/share/7c66921cdbfc4e178f9e6a1a587d5053">https://www.loom.com/share/7c66921cdbfc4e178f9e6a1a587d5053</a> Live Deployed Application: <a href="https://ais-pre-7d5qqkd5x7vda5c5bv7xes-679051571315.europe-west1.run.app">https://ais-pre-7d5qqkd5x7vda5c5bv7xes-679051571315.europe-west1.run.app</a> Quick Walkthrough to Try: Open the live demo and click + Upload Recipe → Load Sample Roast Chicken Note (or upload a photo of your own recipe) to watch the agent extract structured ingredients and timed steps. In Cook Mode, adjust the serving count (- / +) and toggle between Original, Metric (g/ml), and Imperial to see the Python fractions engine recalculate ingredient weights and oven temperatures (°F ↔ °C) in real time. Click Start Timer on any step to launch the multi-timer dock, or click Read Step Aloud for hands-free spoken guidance. Visit the Web Sourcing tab to search the live web for any dish or scrape a recipe URL directly. Visit the Python Agent tab to run the live Python 3.10 subprocess sandbox and download culinary_agent.py. Code GitHub Repository: <a href="https://github.com/Toluwaloju0/Cooking_AI">https://github.com/Toluwaloju0/Cooking_AI</a> Key files in the repository: agent/culinary_agent.py — Standalone, zero-dependency Python 3 culinary agent harness implementing Schema.org JSON-LD recipe scraping, fractions.Fraction & density-aware unit conversions, timer/temperature regex extraction, and Gemma 3 inference. server.ts — Full-stack Express + subprocess bridge connecting the frontend workbench to python3 agent/culinary_agent.py and Google GenAI / Gemma endpoints. src/components/CookModeWorkspace.tsx — Interactive 3-column kitchen station (Mise en Place checklist, Guided Step Timeline with TTS & timers, and Gemma Sous-Chef sidebar). src/components/WebRecipeSourcing.tsx — Live web recipe discovery and direct URL scraper interface. src/components/PythonAgentInspector.tsx — In-browser Python subprocess test bench and source code inspector. How I Built It Mise en Place is built around a hybrid architecture that pairs Google's open-weight Gemma 3 model with a custom, open-source Python 3 Agent Harness: Open-Weight Model (gemma-3-27b-it): The core conversational Sous-Chef agent uses Gemma 3 27B Instruction-Tuned (gemma-3-27b-it) via our Python inference bridge (run_gemma_inference in agent/culinary_agent.py). Every time my friend asks a question in Cook Mode, the agent injects a structured snapshot of the kitchen state—the active recipe title, scaled ingredient quantities, current step index, and target unit system—so Gemma 3 gives precise, step-specific culinary chemistry and technique advice. We also paired Gemma with multimodal document/photo extraction, live web search grounding, and neural Text-to-Speech (audio/wav) for hands-free step reading. Open-Source Python 3 Agent Harness (agent/culinary_agent.py): Built purely with Python 3.10's with zero external pip dependencies, ensuring instant portability across any machine or local kitchen device. Exposes four deterministic CLI and programmatic tools: scrape_recipe_url(url): Custom HTMLParser that extracts <script type="application/ld+json"> blocks, traverses @graph nodes to locate Schema.org Recipe objects, and parses ISO-8601 durations (PT1H30M → 1h 30m). scale_and_convert_recipe(recipe, target_servings, unit_system): Uses fractions.Fraction to compute exact kitchen fractions (1/8, 1/3, 3/4) and a curated INGREDIENT_DENSITY_GRAMS_PER_CUP lookup table (120g/cup flour, 227g/cup butter, 195g/cup Arborio rice) for volume-to-weight conversions, plus regex temperature conversion (450°F ↔ 230°C). parse_raw_recipe_text(raw_text): Normalizes Unicode fractions (½, ¾, ⅓), tokenizes ingredient lines into amount/unit/notes/category, and extracts step timer durations in seconds. run_gemma_inference(prompt, system_instruction, preferred_model): Invokes gemma-3-27b-it and returns transparent execution telemetry (Python version, model used, and execution latency in milliseconds). Full-Stack Kitchen Workbench: Built with React 19, TypeScript, Tailwind CSS, and an Express server (server.ts) that spawns python3 agent/culinary_agent.py via child_process.execFile and returns live execution traces to the UI. Why Does Open Innovation Matter? Cooking is deeply personal, cultural, and domestic—which makes open-source and open-weight AI essential to what Mise en Place achieves: Deterministic Culinary Math Where Black-Box Models Hallucinate: Baking and cooking rely on unforgiving chemical ratios (such as a 78% hydration sourdough boule or a 2% salt cure). Relying solely on a closed, black-box chat prompt to scale recipes or convert volume to weight frequently leads to hallucinated arithmetic and ruined meals. By building an open, inspectable Python tool harness (culinary_agent.py) alongside an open-weight Gemma 3 model, we separate concerns cleanly: deterministic Python fractions and density tables handle exact measurement math and Schema.org HTML scraping, while Gemma 3 handles culinary intuition, flavor pairings, and ingredient substitutions. Privacy for Family Heirlooms & Local/Offline Kitchen Autonomy: Many of the recipes my friend wanted to digitize are treasured family heirlooms that shouldn't be locked inside a proprietary cloud subscription. Because our agent is powered by open-weight Gemma 3 and a zero-dependency standard-library Python script, the entire culinary_agent.py harness can be downloaded with one click from the app and pointed at a local Gemma 3 instance (via Ollama, vLLM, or Hugging Face Transformers on a laptop or Raspberry Pi kitchen counter display) with zero cloud lock-in. Community Extensibility for Global Cuisines: A closed API cannot easily be customized for hyper-local ingredients or regional kitchen measurements. Because our Python agent harness is completely open and inspectable in the Python Agent tab, any home cook or developer can add regional ingredient densities (e.g., cassava flour, palm oil, or local grains), custom unit aliases, or dietary rules in seconds.

My sister needs help cooking meals in the kitchen. Agents which read out recipes and provide feedback and improvements should assist her

Toluwaloju Kayode — Sun, 04 Oct 2026 21:10:05 +0000

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend

What I Built

I built Mise en Place, an AI-powered culinary sous-chef and interactive recipe workbench designed for a close friend who recently moved into their first apartment and wanted to stop relying on expensive takeout and start cooking real meals at home—especially cherished family recipes and authentic regional dishes they find online.

Whenever my friend tried to cook, three real-world kitchen frustrations consistently got in their way:

Scattered, Unusable Recipe Formats: Family recipes were trapped in photos of handwritten index cards or messy text messages, while online recipes were buried under 15 paragraphs of blog backstory, pop-ups, and ads.
Kitchen Math & Scaling Anxiety: Cooking for one or two people when a recipe is written for six meant doing fraction arithmetic (1/3 of 1 ¾ cups) and converting US volume cups into metric grams while food was actively sizzling on the stove.
Mid-Cook Panic with Messy Hands: Halfway through cooking, they would realize they were missing a key ingredient (like mirin or dry white wine) or didn't know what a sensory cue like "cook until translucent" or "al dente all'onda" actually looked like—and with flour- or oil-covered hands, scrolling on a phone screen was a nightmare.

Mise en Place solves all three problems by turning any recipe into a calm, hands-free kitchen station:

Multi-Format Recipe Upload (+ Upload Recipe): Upload a photo/screenshot of a cookbook page or handwritten recipe card, select a .txt/.md/.json/.pdf recipe file, or paste raw kitchen notes. The agent extracts categorized ingredients, exact numeric quantities, temperatures, and step-by-step instructions with embedded timer durations.
Live Web Recipe Sourcing & Python Schema.org Scraper (Web Sourcing): Search the web by dish name or by the ingredients currently sitting in your fridge, or paste any recipe URL. Our Python agent (scrape_recipe_url) parses <script type="application/ld+json"> Schema.org Recipe graphs directly from webpage HTML to strip away blog clutter while preserving clickable source citations.
Deterministic Python Fraction & Density Scaling (Cook Mode): Scale any recipe from 1 to 24 servings and toggle between Original, Metric (g / ml / °C), and Imperial (oz / cup / °F) with real-time oven temperature conversions inside the step text.
Smart Multi-Timer Dock & Hands-Free Voice Read-Aloud: Every time-bound cooking step generates a one-click countdown timer with Web Audio API chimes so multiple stove and oven stages can run simultaneously, plus a Read Step Aloud voice button for hands-free listening when your hands are covered in dough.
Contextual Gemma 3 Sous-Chef Sidebar: Ask real-time culinary questions ("What can I substitute for white miso?", "What sensory cues tell me this step is done?", "How can I prep this dish ahead of time?") with full awareness of your active recipe, current step number, and scaled serving size.

Demo

https://www.loom.com/share/7c66921cdbfc4e178f9e6a1a587d5053

Live Deployed Application: https://ais-pre-7d5qqkd5x7vda5c5bv7xes-679051571315.europe-west1.run.app
Quick Walkthrough to Try:
1. Open the live demo and click + Upload Recipe → Load Sample Roast Chicken Note (or upload a photo of your own recipe) to watch the agent extract structured ingredients and timed steps.
2. In Cook Mode, adjust the serving count (- / +) and toggle between Original, Metric (g/ml), and Imperial to see the Python fractions engine recalculate ingredient weights and oven temperatures (°F ↔ °C) in real time.
3. Click Start Timer on any step to launch the multi-timer dock, or click Read Step Aloud for hands-free spoken guidance.
4. Visit the Web Sourcing tab to search the live web for any dish or scrape a recipe URL directly.
5. Visit the Python Agent tab to run the live Python 3.10 subprocess sandbox and download culinary_agent.py.

Code

GitHub Repository: https://github.com/Toluwaloju0/Cooking_AI

Key files in the repository:

agent/culinary_agent.py — Standalone, zero-dependency Python 3 culinary agent harness implementing Schema.org JSON-LD recipe scraping, fractions.Fraction & density-aware unit conversions, timer/temperature regex extraction, and Gemma 3 inference.
server.ts — Full-stack Express + subprocess bridge connecting the frontend workbench to python3 agent/culinary_agent.py and Google GenAI / Gemma endpoints.
src/components/CookModeWorkspace.tsx — Interactive 3-column kitchen station (Mise en Place checklist, Guided Step Timeline with TTS & timers, and Gemma Sous-Chef sidebar).
src/components/WebRecipeSourcing.tsx — Live web recipe discovery and direct URL scraper interface.
src/components/PythonAgentInspector.tsx — In-browser Python subprocess test bench and source code inspector.

How I Built It

Mise en Place is built around a hybrid architecture that pairs Google's open-weight Gemma 3 model with a custom, open-source Python 3 Agent Harness:

Open-Weight Model (gemma-3-27b-it):
- The core conversational Sous-Chef agent uses Gemma 3 27B Instruction-Tuned (gemma-3-27b-it) via our Python inference bridge (run_gemma_inference in agent/culinary_agent.py).
- Every time my friend asks a question in Cook Mode, the agent injects a structured snapshot of the kitchen state—the active recipe title, scaled ingredient quantities, current step index, and target unit system—so Gemma 3 gives precise, step-specific culinary chemistry and technique advice.
- We also paired Gemma with multimodal document/photo extraction, live web search grounding, and neural Text-to-Speech (audio/wav) for hands-free step reading.
Open-Source Python 3 Agent Harness (agent/culinary_agent.py):
- Built purely with Python 3.10's with zero external pip dependencies, ensuring instant portability across any machine or local kitchen device.
- Exposes four deterministic CLI and programmatic tools:
- scrape_recipe_url(url): Custom HTMLParser that extracts <script type="application/ld+json"> blocks, traverses @graph nodes to locate Schema.org Recipe objects, and parses ISO-8601 durations (PT1H30M → 1h 30m).
- scale_and_convert_recipe(recipe, target_servings, unit_system): Uses fractions.Fraction to compute exact kitchen fractions (1/8, 1/3, 3/4) and a curated INGREDIENT_DENSITY_GRAMS_PER_CUP lookup table (120g/cup flour, 227g/cup butter, 195g/cup Arborio rice) for volume-to-weight conversions, plus regex temperature conversion (450°F ↔ 230°C).
- parse_raw_recipe_text(raw_text): Normalizes Unicode fractions (½, ¾, ⅓), tokenizes ingredient lines into amount/unit/notes/category, and extracts step timer durations in seconds.
- run_gemma_inference(prompt, system_instruction, preferred_model): Invokes gemma-3-27b-it and returns transparent execution telemetry (Python version, model used, and execution latency in milliseconds).
Full-Stack Kitchen Workbench:
- Built with React 19, TypeScript, Tailwind CSS, and an Express server (server.ts) that spawns python3 agent/culinary_agent.py via child_process.execFile and returns live execution traces to the UI.

Why Does Open Innovation Matter?

Cooking is deeply personal, cultural, and domestic—which makes open-source and open-weight AI essential to what Mise en Place achieves:

Deterministic Culinary Math Where Black-Box Models Hallucinate: Baking and cooking rely on unforgiving chemical ratios (such as a 78% hydration sourdough boule or a 2% salt cure). Relying solely on a closed, black-box chat prompt to scale recipes or convert volume to weight frequently leads to hallucinated arithmetic and ruined meals. By building an open, inspectable Python tool harness (culinary_agent.py) alongside an open-weight Gemma 3 model, we separate concerns cleanly: deterministic Python fractions and density tables handle exact measurement math and Schema.org HTML scraping, while Gemma 3 handles culinary intuition, flavor pairings, and ingredient substitutions.
Privacy for Family Heirlooms & Local/Offline Kitchen Autonomy: Many of the recipes my friend wanted to digitize are treasured family heirlooms that shouldn't be locked inside a proprietary cloud subscription. Because our agent is powered by open-weight Gemma 3 and a zero-dependency standard-library Python script, the entire culinary_agent.py harness can be downloaded with one click from the app and pointed at a local Gemma 3 instance (via Ollama, vLLM, or Hugging Face Transformers on a laptop or Raspberry Pi kitchen counter display) with zero cloud lock-in.
Community Extensibility for Global Cuisines: A closed API cannot easily be customized for hyper-local ingredients or regional kitchen measurements. Because our Python agent harness is completely open and inspectable in the Python Agent tab, any home cook or developer can add regional ingredient densities (e.g., cassava flour, palm oil, or local grains), custom unit aliases, or dietary rules in seconds.

From APIs to AI Agents: How My Skills Apply to Today’s Software World

Toluwaloju Kayode — Fri, 02 Oct 2026 21:48:30 +0000

I’ve spent the past few years learning by building.

Rather than limiting myself to one area of software development, I’ve explored backend engineering, frontend development, APIs, databases, cloud deployment, automation, AI agents, and cybersecurity.

What I’ve realized is that these skills are becoming increasingly interconnected.

Today, building a useful application isn't just about writing code. It can involve designing an API, creating a user interface, connecting databases, deploying to the cloud, automating repetitive processes, and increasingly, integrating AI into the workflow.

That is the direction I’m currently taking as an AI and Full-Stack Developer.

**🧑🏽‍💻 My Core Technical Skills

**
My current toolkit covers several areas of modern software development:

🐍 Python

Python has been one of the most important parts of my development journey.

I've used it for:

Backend development
REST APIs
Automation
Data processing
AI applications
Scripting Frameworks such as FastAPI and Django allow me to turn Python knowledge into production-oriented applications.

⚡ JavaScript & Full-Stack Development

I've also worked extensively with JavaScript and modern frontend technologies.

My experience includes:

React
Next.js
Vite
REST API integration
Responsive interfaces
Real-time communication

This allows me to work across both the frontend and backend rather than treating them as completely separate worlds.

🔌 APIs & Backend Systems

A large part of modern software is essentially about systems communicating with other systems.

I've worked with:

REST APIs
Authentication
JWT
OAuth/Google authentication
WebSockets
Microservice-style architectures
API integrations

Building APIs has also helped me understand how frontend applications, databases, third-party services and AI systems can communicate.

🗄️ Databases

I've worked with both relational and NoSQL databases, including:

PostgreSQL
MySQL
MongoDB
SQL
Database modeling and queries

Understanding databases is particularly important when building applications that need to manage real-world information rather than simply displaying static content.

☁️ Cloud, DevOps & Deployment

Writing an application is only part of the job.

I've also worked with technologies and platforms such as:

Docker
Kubernetes
Nginx
AWS
Vercel
Render
Railway
Linux
CI/CD

Working through deployment issues has taught me something that tutorials don't always show: software behaves differently when real infrastructure gets involved.

🤖 Moving Into AI Agents

One of the biggest areas I'm currently exploring is AI agent development.

I'm particularly interested in how AI can move beyond simple question-and-answer interactions and actually participate in workflows.

I'm exploring concepts such as:

AI agents
Tool calling
RAG
LangChain
LangGraph
Automation
Agentic workflows
AI-powered APIs

The combination of traditional software engineering and AI is especially interesting to me.

An AI agent still needs many things that normal software needs:

APIs.
Databases.
Authentication.
Business logic.
Security.
Deployment.
Monitoring.

This is where my full-stack background becomes useful.

🔐 Why I'm Exploring Cybersecurity

Another area I've recently been developing is cybersecurity.

I started learning more about cybersecurity operations and ethical hacking while working on an idea for a cybersecurity-focused AI agent.

The goal isn't simply to put "AI" into cybersecurity.

I'm interested in exploring practical questions such as:

Can AI help automate security workflows?
Can agents analyze security information?
Can repetitive security tasks be automated?
How can AI assist developers in identifying potential risks?
How can software engineering and cybersecurity work together?

I'm still learning the cybersecurity side, but that's actually part of what makes the project interesting.

I'm learning the domain while simultaneously thinking about how software can be applied to it.

🏗️ Turning Skills Into Real Projects

The most important part of my learning has been building projects instead of only collecting technologies.

🏥 MediConnect

I've also been working on MediConnect, a clinic management and appointment-booking platform.

The project focuses on reducing unnecessary waiting time by allowing patients to discover doctors and view availability.

It involves concepts such as:

Patient and doctor portals
Authentication
Google Sign-In
Appointment management
Doctor schedules
Backend APIs
Database management

This project demonstrates another important part of software development:

technology should solve a real problem.

🛡️ Cybersecurity AI Agent

My cybersecurity AI project combines several areas I'm currently learning:

Software Engineering + AI + Automation + Cybersecurity

The goal is to explore how an AI agent could assist with cybersecurity-related workflows.

This project has also pushed me outside my usual software development comfort zone and encouraged me to learn more about cybersecurity operations.

🚀 What I'm Building Toward

My current direction is simple:

Build → Learn → Experiment → Improve → Build again.

I'm working toward becoming stronger at building complete systems that combine:

Full-Stack Development
+
Backend Engineering
+
AI Agents
+
Automation
+
Cloud Infrastructure
+
Cybersecurity

I don't expect to know everything.

Technology changes too quickly for that.

Instead, I want to remain the kind of developer who can encounter a new problem, learn what is necessary, experiment with the technology, and turn the knowledge into something useful.