Summary
Andrey Buzin is a software engineer with eight years of experience building AI products, currently based in San Francisco and working at Vercel. He has co-founded and led ML engineering efforts at startups, shipping data-generation agents, multimodal LLMs, and a fast parallel evaluation system that improved throughput tenfold on limited hardware. His background spans applied research to production: from cross-camera tracking and anti-spoofing optimizations to industrial safety and multi-sensor vehicle monitoring, often cutting pipeline times and memory footprints dramatically. Known for pragmatic tooling, he created frameworks that reduced VLM training setup from days to hours and slashed hallucinations in RAG systems to near zero on curated benchmarks. Beyond models, he focuses on reproducible ML engineering—integrating Airflow, artifact registries, and experiment tracking to increase experiment throughput and free up dozens of hours per week.
8 years of coding experience
6 years of employment as a software developer
Master's degree Applied Mathematics and Computer Science, Master's degree Applied Mathematics and Computer Science at Saint Petersburg State University
English, German, Russian