Fullstack Developer · AI Integrator · Automation Engineer · System Problem Solver
I’m a fullstack developer with a passion for solving real-world problems using code, architecture, and AI-powered systems.
My journey so far has been all about building, automating, and improving — not just completing tasks. I love working on systems that directly affect users, improve developer experience, and scale as the business grows.
Even though many of the systems I’ve built are internal and currently private, here’s what I’ve been working on:
- Integrated LLM models (Gemini, Qwen, Hugging Face) into custom backend systems via prompt engineering.
- Built dynamic AI prompting engine: changing model behavior based on category, use-case, or query param.
- Developed semantic prompt-to-SQL system: users can search for candidates with natural language — no UI filter, just prompt → auto SQL → results.
- Built speech-to-text transcription with speaker diarization, including async handling with FastAPI and Socket.IO.
- Converted face recognition & liveness models (ArcFace HF5, FasNet) to ONNX for production-grade performance.
- Migrated full legacy system into Laravel 10, improving maintainability and long-term scalability.
- Developed microservices with FastAPI and Flask, integrated with RabbitMQ and Redis.
- Created RabbitMQ task decorator in Python to simplify background job creation and improve code consistency.
- Designed and documented RESTful and real-time APIs with Socket.IO integration for AI chat and task feedback.
- Engineered multi-language Docker environments (Laravel + FastAPI + Redis + RabbitMQ) to support AI workflows and multi-service deployment.
- Fixed critical Docker env/config errors to support both dev & prod pipelines with consistency.
- Setup Redis & RabbitMQ on new servers with secure config & volume persistence.
- Diagnosed and helped resolve PHP-FPM pool exhaustion and race condition bugs in job processing.
- Built internal tool for LinkedIn People scraping using headless browsers and stealth mode to enrich internal candidate search.
- Collaborated with Mobile team to enforce face image cropping standards, improving face recognition accuracy across platforms.
- Performed Windows OS-level recovery for a locked device without reinstall — using deep system utility access.
- Prototyped a recommendation system for content posts based on semantic matching and user behavior.
- Regularly act as bridge between backend, mobile, and AI teams — not just writing code, but proposing structure, improving systems, and reducing blockers.
Languages: PHP, Python, JavaScript
Frameworks: Laravel, FastAPI, Flask
AI Tools: Hugging Face, Gemini, ONNX, OpenAI, ArcFace
Infra & DevOps: Docker, Redis, RabbitMQ, Nginx
Automation: Selenium, Puppeteer, Undetected-Chromedriver
Databases: MySQL, PostgreSQL
Others: Socket.IO, REST API, WebSocket, CI/CD, Linux
I don't chase hype — I chase what works.
I believe in clean, scalable, and meaningful systems.
If it's slow, I speed it up.
If it's unclear, I refactor it.
If it's repetitive, I automate it.
If it's a wall, I climb it — or build a smarter path.
Most of my contributions are in private repositories or internal projects under NDA, which means my GitHub stats might not reflect the systems I’ve helped build.
But behind the scenes, I’ve delivered:
- Production AI features
- High-traffic backend services
- Critical infrastructure fixes
- Dev experience improvements
- And real collaboration with teams across the stack
If you want to collaborate, connect, or just share ideas — I’m always open to build something meaningful.
- 📧 Email: azzarnuzy93@gmail.com
- 💻 GitHub: github.com/Azzarnuji