Senior Engineer building production LLM tooling — grounding, evaluation, and agentic systems. I turn LLM capabilities into reliable, deployable software.
⚓ anchor — fact-check for AI outputs
A runtime grounding gate that catches hallucinated claims with exact citations. 0.93 F1 on FEVER (public benchmark). Fully local, zero API tokens. Live demo
Natural-language to SQL queries with an interactive dashboard.
Python PyTorch Transformers NLI RAG FastAPI Docker pandas SQL
Every project ships with an eval.



