Robotics systems
ROS 2 integration, navigation, manipulation, control scaffolds, and sim-first validation.

I’m Obinna Edeh, an AI systems engineer building dependable robotics, edge inference, digital-twin, and operational intelligence systems.
I build operational AI systems for physical and edge environments under real infrastructure constraints.
My background leading nationwide wireless deployment gives me a systems-level view of AI deployment, especially where telemetry, infrastructure reliability, edge inference, and real-time operational constraints matter.
My current work spans Physical AI, edge inference, computer vision, runtime observability, and operational safety. Across my flagship projects, I work with Jetson-based inference, ROS 2 robotics, simulation, defensive telemetry, and time-series forecasting, with AI-RAN KPIs as one telemetry source among others.
The common thread is simple: AI systems should not just run once in a demo. They should be measurable, observable, reliable, and useful under real constraints. My passion is bridging the sim-to-real gap.
Each project is built around reproducible artifacts, operational evidence, and a clear path to deployment.
Robot manipulation program spanning measured Jetson inference and probes, measured Isaac Sim evaluation, and an implemented but unmeasured Orin ROS 2 host.
Runtime safety layer measured on Jetson AGX Thor on 2026-09-09: 28.6 frames/s pipeline overhead with a mock model, and Cosmos-Reason2-2B inference at p50 4.35 s per frame. Detection quality unmeasured; no ground truth yet.
Defensive edge telemetry runtime with measured Jetson AGX Thor inference and thread-pool power, plus operator-reviewed alerts.
ROS 2 integration, navigation, manipulation, control scaffolds, and sim-first validation.
Jetson inference and edge runtime workflows measured with latency, telemetry, reliability, power, and thermal evidence.
Isaac Sim and OpenUSD workcells that connect robot assets to repeatable validation.
Human-bounded copilots that reason over telemetry, logs, manuals, and runtime state.