Lucas Butler

I’m a full-stack AI software engineer who builds production-grade end-to-end systems—pairing directly with customers to pinpoint pain points and ship solutions that hold up in real, high-trust environments. At Sandia National Laboratories, I turned ambiguous workflows into an LLM agent pipeline deployed across 5 teams, saving $20,000/year per team, while also engineering it to run fully air-gapped on a classified network with zero external dependencies.

I combine software engineering rigor with practical AI engineering. I built a resume screening pipeline that processes 100+ resumes/minute at peak throughput using Flask/PostgreSQL and secured $5,000 in external funding by presenting the design and results to stakeholders; I’ve also documented and iterated with engineering teams to drive adoption.

I’m equally energized by applied R&D and research that translates into tools. I designed a multi-service ADAS system on Jetson Orin Nano (25+ FPS) and trained YOLO11n for traffic light detection (90% accuracy), and I co-authored an SPIE paper on CNNs vs RNNs for elephant infrasound detection (95%). My ethos is simple: prototype fast, validate thoroughly, and deliver systems that are maintainable, testable, and ready for deployment.