AI/ML Engineer · MSc Artificial Intelligence · 3+ years industry experience
I specialise in Agentic RAG systems, LLM pipelines, and production AI infrastructure — with a focus on evaluation, reliability, and deployment under real constraints.
| Project | What it does | Stack |
|---|---|---|
| DocRAG | Agentic RAG system with a ReAct agent routing between hybrid retrieval (BM25 + ChromaDB) and Wikipedia — 0.90 context precision on RAGAS eval | LangGraph · GPT-4o · ChromaDB · BM25 |
| Production Agentic AI Pipeline | Multi-agent LangGraph system with LangSmith tracing and a full RAGAS eval loop — faithfulness 1.0, answer relevancy 0.932; eval-driven fix doubled context recall | LangGraph · GPT-4o · LangSmith · RAGAS · Streamlit |
| AI Travel Itinerary Agent | ReAct agent that plans day-by-day trips using real-time web search | LangGraph · Claude Sonnet · Kubernetes · ELK |
| Real-time AI Gym Coach | Webcam-based AI trainer with pose detection and live voice coaching — 92% form accuracy | MediaPipe · Claude AI · Streamlit · SQLite |
LLMs & Agents LangChain · LangGraph · Claude API · OpenAI · Prompt Engineering
RAG & Search ChromaDB · BM25 · Hybrid Search · Semantic Chunking · RAGAS
AI/ML PyTorch · MediaPipe · Computer Vision · Deep Learning
Infrastructure Docker · Kubernetes · ELK Stack · GitHub Actions · CI/CD
Backend Python · FastAPI · SQLite · Streamlit
- MSc Artificial Intelligence @ BTU Cottbus (expected 2026)
- Building production-grade agentic RAG systems
- Open to full-time roles and research collaborations in Generative AI / LLM Engineering