Perception Engineer

We are looking for a Perception Engineer with 4-7 years of experience to own and evolve the perception stack powering Sauron's autonomous home security platform. This is a product-first role — Sauron already has installs in the field and active beta programs, so you'll be shipping real improvements to a deployed system running on NVIDIA edge hardware, not doing research in a lab.

What will you be doing?

• Re-architecting and optimizing the existing perception service running on NVIDIA Thor to support more camera streams and improve pipeline stability
• Evaluating and swapping in faster, higher-performing models while maintaining real-time performance on edge hardware
• Bringing monocular depth estimation onto the NVIDIA Thor platform to enable 3D tracking of people and vehicles around properties
• Working deeply with NVIDIA DeepStream and the Jetson framework to squeeze maximum performance out of the edge compute stack
• Collaborating cross-functionally with the device team and systems engineers to ship perception improvements to customers in the field

Key Requirements

• Strong proficiency in Python and C++ with hands-on experience building production perception or ML systems (not research-only)
• Real experience with the NVIDIA stack — DeepStream, GStreamer, TensorRT, and ideally Jetson (Thor/Orin) edge deployment
• Background in computer vision for hardware products in robotics, autonomous vehicles, aerospace, or similar domains
• Understanding of video pipelines (GStreamer, FFmpeg) and how camera feeds actually work in production
• Self-directed IC who can take ownership, unblock themselves, and work cross-functionally in a fast-moving startup environment