Perception Engineer, Edge Computer Vision (Home Security AI)

A startup building an autonomous home security system, with units already installed and beta programs running, needs an engineer to take charge of its on-device vision software. The job is about shipping measurable upgrades to a product that is already deployed on embedded GPU hardware, not lab research.

Your focus

• Rework and tune today’s perception service so the device handles additional camera feeds and runs more stably
• Test faster, stronger models and swap them in without losing responsiveness on embedded hardware
• Add monocular depth estimation to the device so people and cars can be tracked in three dimensions near a property
• Get maximum performance from the DeepStream and Jetson tooling NVIDIA provides
• Work with the device and systems teams to deliver perception upgrades to customers out in the world

What you bring

• Strong Python and C++ plus production-grade perception or ML work, not only academic projects
• Hands-on time with NVIDIA tooling: DeepStream, GStreamer, TensorRT, ideally Jetson Orin or Thor deployments
• Computer vision background for physical devices in domains like aerospace, robotics or self-driving
• Practical knowledge of video pipelines and live camera feeds, including GStreamer and FFmpeg
• Roughly three to eight years of experience, a self-directed style and comfort with startup pace

Hybrid in the San Francisco Bay Area. Visa transfers considered, though L1 transfers cannot be supported.