Selected work

The areas I work in — kept general, since most of the detail lives in internal systems.

My recent work is at Tactus AI, on robotics and ML for diagnostic lab automation. A high-level look at the kinds of problems I spend my time on.

01Agentic operations

Using language models to help operate infrastructure — handling routine issues automatically, with evaluation and guardrails so autonomy is added gradually and safely.

02ML & GPU inference infrastructure

Serving and evaluating models in production, self-hosted where privacy or cost calls for it, with a focus on keeping inference dependable and affordable.

03Platform architecture

Helping shape the architecture a growing engineering team builds on — clear boundaries between components and contracts that let teams move independently.

04Autonomy — navigation & SLAM

Navigation and localization for mobile robots: planning, mapping, and the sensor and estimation work that turns a demo into something dependable.

05Embodied AI & control

Real-time control and learned manipulation for robot arms, with the safety engineering that has to sit underneath any of it.

06Infrastructure as code & reliability

Provisioning, secrets, and recovery managed as reviewed, auditable code, plus the ongoing reliability work that keeps a fleet running.

07Regulated software

Software for regulated, medical-device settings, and the operator-facing tools people use day to day.

08Career