About Me

Messy data in. Production AI out.

I help teams turn trapped document data into automated, production-grade AI systems. For 8+ years I've taken the messiest inputs (handwritten medical forms, crumpled scans, insurance documents) and delivered measurable results: an 88%-accurate extraction pipeline, agentic chatbots that replace hours of manual cross-referencing, and self-hosted LLMs that cut inference costs while keeping sensitive data in-house. 10+ ML products shipped to production, not notebooks. I lead a GenAI team and specialize in taking models from prototype to Kubernetes without breaking prod. AI Engineer, Data Scientist, and Software & AI Architect, working with PyTorch, LLMs, and more YAML than I'd like to admit. Since Feb 2026 I'm also part-time Software & AI Architect at IngeniousZone, building the CI/CD rails that AI products ride into production.