Computer Science Engineering · Applied AI · Information Retrieval · Full-Stack Systems
I build systems that combine experimentation with practical engineering — from retrieval pipelines and browser-side ML to production-style backend APIs, deployment workflows, and user-facing applications.
Experimental hybrid retrieval-augmented generation system evaluating quantum-simulated retrieval and reranking against classical baselines.
Highlights
- BM25 + dense retrieval baselines
- Quantum-kernel experiments
- Grover-inspired amplification
- QAOA reranking
- Ablations, statistical testing, reproducibility
Stack
Python FastAPI Qiskit FAISS BM25 NumPy pytest
Full-stack computerized maintenance management system for QR-based equipment workflows, predictive alerts, reporting, and mobile-friendly maintenance operations.
Highlights
- QR-based machine workflows
- Predictive maintenance alerts
- Inventory burn-rate monitoring
- Automated reporting
- Mobile-oriented technician interface
Stack
Next.js React Flask SQLite APScheduler Nginx DigitalOcean
Browser-side waste classifier using ONNX Runtime Web so image inference runs locally instead of requiring a remote inference server.
Highlights
- Client-side ONNX inference
- Camera and image upload support
- Local processing for privacy
- Responsive browser UI
Stack
ONNX Runtime Web PyTorch JavaScript Tailwind CSS
Open, dependency-free coordination kit for AI-assisted software work across Codex, Claude Code, GitHub Copilot, Antigravity, Ollama, and Hermes Agent—with native NVIDIA Nemotron 3 Ultra routing.
Highlights
- Routes work by task fit, privacy, cost, latency, and available tools
- Prevents parallel agents from claiming the same files
- Creates portable handoffs with verification evidence and risk notes
- Includes local-private routing and automated secret-leak protection
Stack
Python GitHub Actions Ollama Codex Claude Code GitHub Copilot Hermes Agent NVIDIA Nemotron
Happiness Project
Privacy-aware Flask journaling platform with authentication, private/public/shared posts, and SQLAlchemy relationships.
Birkbeck Fire Drill
Three.js first-person browser simulation featuring hazards, oxygen mechanics, WebGL rendering, and interactive game-state logic.
Python Java JavaScript TypeScript SQL
SQLite Git Linux Nginx DigitalOcean
| Area | Work |
|---|---|
| Research | Quantum-enhanced retrieval-augmented generation, experimental evaluation, ablations and reproducibility |
| Natural Language Processing | Academic coursework and applied NLP assignments |
| Multimedia Technologies | Technical report work involving multimedia systems and concepts |
| International / SAP Coursework | Country-analysis and international academic work |
| Networking | CCNA: Introduction to Networks certification |
| Data Science | Python for Data Science certification |
| Industry Experience | Industrial maintenance / CMMS internship project with full technical documentation |
- Build systems I can explain end-to-end
- Measure results instead of assuming an approach works
- Prefer reproducible experiments and clear documentation
- Treat security and deployment hygiene as engineering requirements
- Learn the underlying computer-science concepts, not only the frameworks
Stats are generated from public GitHub repository data and may not represent every language or contribution.