With over 6 years of experience in distributed systems, I design and scale high-performance architectures — focusing on vectorized data processing, real-time systems, and applied AI.
Most recently, I re-architected a legacy monolith into a distributed Polars & Dagster pipeline — processing 40M+ daily records and reducing ETL runtime by 97% (5.5h → 16m).
I recently architected a company-wide MCP server enabling plain-English search across product data — achieving 45% prompt-to-pageview conversion in its first week, greenlit for public release.
Previously, I engineered Zoom for Government infrastructure (FedRAMP/IL4) to support federal agencies.
Before that, I developed low-latency voice processing systems and DSP encryption in C++ at Motorola for mission-critical 911 dispatch communications.