I like problems that are half research, half plumbing.

Founding AI-focused full-stack engineer with 4+ years shipping production systems for large user bases. I build reliable AI products end to end — LLM workflows, retrieval pipelines and real-time backend architecture — while balancing performance, quality and cost.

I also design them. Most AI products fail at the interface rather than the model, so empty states, latency affordances and information density get as much of my attention as retrieval quality does.