NVIDIA

Fleet Command was built for enterprises deploying edge AI at scale. The companies writing that software were startups. Teams of five, running computer vision and MLOps in hospitals, factories, and shops. They did not want a procurement conversation.

For them, the gap between "this could run my workload" and "my workload is running" was measured in weeks.

I led developer relations for EGX, Fleet Command, and the Jetson IoT and Android orchestration software ecosystem.

I built a self-serve path. Helm chart templates put Kubernetes workloads straight onto NGC Cloud, so a startup could start on its own. An onboarding motion picked up the 50+ that wanted help, across medical, industrial, robotics, and retail.

I also read more than eighty cloud-native platforms, and worked with Jetson software engineering to enable 15+ orchestration platforms on the hardware.

Reading eighty platforms took six weeks. I wrote it up as a position on where Fleet Command had to go. That is what moved product and security decisions.

Developer relations talks to more teams running the product under real load than anyone else in the building. Next to the roadmap, that knowledge gets used. Next to marketing, it gets thrown away every quarter.

50+ edge AI startups onboarded onto Fleet Command, building fleet management and IoT solutions on Jetson and EGX. 50+ Helm chart templates for self-serve deploy to NGC Cloud. 80+ cloud-native platforms turned into a position that moved the roadmap. 15+ orchestration platforms enabled on Jetson.