ML systems that hold up in production.

I build end-to-end ML pipelines and the backend infra they run on — mostly in telemetry-heavy, fleet-scale domains.

Recent focus: causal inference for driver efficiency, real-time route matching, and LLM tooling for ops.