I test flight-critical rocket hardware for a living. On the RS-25 program I run cryogenic inspections, develop test sequences to flight-qualification standards, and operate the pressurized systems that verify an engine is safe to fly. The job is unforgiving about one thing above all: nothing ships until it is verified.

On nights and weekends I took that same discipline and pointed it at AI/ML infrastructure. Self-taught and self-built, I now design and operate real, production-style systems — an agentic RAG knowledge base that measurably cut research time, a ~32k-LOC local multi-agent platform with an adversarial peer-review loop, and a two-node multi-GPU cluster that serves local models with a privacy-first architecture.

The through-line is reliability engineering: measured-not-asserted numbers, reproducible config-as-truth, least-privilege and fail-closed defaults, and testing before shipping. I am looking for remote-first work building trustworthy, tested AI infrastructure.