I'm a Computer Science Graduate from FAST who got a bit obsessed with how language models actually think — and ended up spending most of my time building agentic systems, RAG pipelines, and AI-powered systems that actually work.

I care about building things end-to-end: from designing the retrieval architecture to shipping a working interface. I'm not a fan of demo-only projects.

AI / ML

Vector DBs & Retrieval

Backend & Infra

Frontend

Agentic-RAG Resume Screener A multi-strategy recruitment system that routes queries across four RAG strategies — Fusion RAG, HyDE, CRAG, and Graph RAG — over 2,400+ resumes. Built a LangGraph orchestrator on top of a hybrid retrieval pipeline to maximize relevance matching.

CoWriteIA AI writing platform with context-aware assistance, semantic RAG, and document indexing. Built the embedding pipeline from scratch and wired it into a FastAPI backend with a Next.js frontend. The goal was a smarter alternative to existing writing tools that actually understands context.

Velora Multi-agent framework for financial, risk, and competitor analysis. Each agent has a defined scope and they talk to each other through self-correction feedback loops using LangChain. Learned a ton about agent coordination and prompt engineering here.

HireSmart ATS-optimized resume builder with Supabase auth and Puppeteer-based PDF export. Built this partly because I was frustrated with how bad most resume tools are — wanted to prove you could do better.

• Deep in LLM orchestration — specifically how to make multi-agent systems reliable, not just impressive in demos
• Exploring evaluation frameworks (RAGAS, custom evals) to actually measure whether a RAG system is working
• Shipping real-world projects with agentic systems

I mentored junior students in AI/ML at university, competed in the International Mathematics Olympiad (and medalled), and I paint when I need a break from staring at embeddings. I believe the best AI systems come from deeply understanding both the theory and the practical constraints of deployment.

Feel free to reach out if you're building something in the agentic/RAG space, or just want to talk shop.