From error-correction research to production AI systems.

I'm an engineer who moves comfortably between research rigor and shipping velocity. My path started in automotive embedded software at Bosch, went deep into deep-learning research during my M.S. at the University of Ulsan — where I published first-author in an IEEE Transactions journal and won a Best Paper Award — and now lives in production AI product engineering at Edge8 AI.

Over the last 18 months I've shipped and maintained roughly 40 production repositories: multi-tenant SaaS backends, from-scratch multi-agent systems, RAG pipelines, streaming AI interfaces, and a clinical risk engine for regulated health-tech. I like owning the whole lifecycle — architecture, implementation, testing, and the documentation that keeps a system legible.

I care about the unglamorous parts: tenant isolation you can prove with a test, retrieval strategies you can benchmark, and LLM systems you can evaluate rather than hope about. My workflow is deliberately AI-augmented — I build the agent tooling I use to build.