Professional Overview
AI specialist who builds and ships production-grade LLM and machine-learning systems end to end — from data pipelines and model development through evaluation, deployment, and monitoring. I turn research into tools that run reliably in production, not notebook demos. Core stack: Python and PyTorch, with LLM application work in LangChain and RAG pipelines over vector databases such as pgvector and Pinecone; MLOps for CI/CD, tracing, and drift monitoring; and cloud deployment on AWS and GCP. Recent work includes Atlas Support Copilot — a RAG assistant over 12k internal documents that cut first-response handling time ~40% — and Nimbus Forecast, a demand-forecasting model that reduced stockouts 18%. I care about grounding, evaluation, latency, and cost, and I document as I go so teams can pick up the work. Available for full-time, part-time, or contract engagements, remote-first, with reliable overlap for US and EU teams.
