AI ML systems engineer · engineering automation · technical leader
About
I design AI/ML systems and engineering workflows that connect autonomy, embedded intelligence, and guidance and control to reliable fielded performance.
I build AI-enabled engineering systems that work beyond the lab. My career spans guidance, navigation and control, simulation, data and MLOps infrastructure, embedded systems, computer vision, machine learning, autonomy, agentic automation, and technical leadership. I’m most energized by problems where an algorithm, simulation, or data workflow cannot live comfortably in a notebook. It has to fit the engineering loop: execute on target hardware or established tools, respect timing and memory limits, connect to credible data, and survive rigorous verification.
My current work centers on real-time AI/ML and agentic engineering automation for embedded and fielded applications: neural architecture search, infrared target detection and classification, optimized inference, SIMD kernels, software-in-the-loop integration, flight-test data pipelines, and reinforcement-learning research. Earlier roles gave me a broader systems foundation in 6-DOF simulation, digital twins, CI/CD, embedded C/C++, robotics, test systems, and spacecraft guidance and control.
I also hold a Ph.D. in Computer Engineering from the University of New Mexico, where my research focused on safety-constrained decentralized multi-UAV coordination with deep reinforcement learning. That research sharpened an idea that now shapes much of my engineering: intelligent behavior is only useful when it is paired with constraints, observability, safety, and a credible path to implementation.
The problems I like
I gravitate toward engineering questions that force several disciplines to negotiate with each other:
- How do we improve model accuracy without blowing the latency or memory budget?
- How do we connect data, simulations, and test evidence into an engineering workflow that shortens iteration without weakening traceability?
- How do we translate an autonomy concept from simulation into a testable system architecture?
- How do we measure the right thing when model metrics and system metrics disagree?
- How do we build tooling that lets engineers iterate faster without weakening verification?
- How do we make advanced algorithms understandable enough for a broader engineering team to adopt them?
How I work
I tend to think in layers: mission objective → data and system behavior → model or simulation → software architecture → compute → verification evidence. That systems view helps me move between research, implementation, workflow automation, performance optimization, and technical leadership without losing the thread connecting them.
Outside a single technical niche, the recurring theme of my career is simple: build capable systems, understand their constraints, measure them honestly, and improve the bottleneck that matters.