About
Sergei Notevskii
AI Platform Lead building production-grade LLM, STT, embeddings and agent platforms. The public work here is about production taste: making AI systems measurable, operable, cost-aware and useful in real products.
Hard engineering
Self-hosted inference, vLLM, GPU capacity, model routing, cache and latency.
Quality systems
Evals, scenario datasets, feedback loops, regression checks and model release gates.
Platform leadership
Operating model, ownership, DevEx, cost review, incident process and sanitized public frameworks.