About the company
Our client is a technology company.
The role
Raydar is recruiting for this role on behalf of our client. Own and improve a production computer vision service running on embedded hardware, delivering performance and stability gains to a system already in use in the field. The work centers on tuning real-time video and model pipelines and partnering closely with hardware and systems colleagues.
What you'll do
- Re-architect and optimize an existing real-time perception service so it handles more video streams with better stability.
- Evaluate and integrate faster, higher-performing models while keeping real-time performance on edge hardware.
- Bring depth estimation capabilities onto embedded compute to support three-dimensional tracking.
- Tune the video and inference stack on edge devices to get the most out of the hardware.
- Work with hardware and systems engineers to ship perception improvements to deployed systems.
Requirements
What we're looking for
- 4 to 7 years of experience, with strong Python and C++ skills.
- Hands-on experience building production perception or machine learning systems, beyond research work.
- Familiarity with GPU-accelerated inference tooling and deployment on embedded edge devices.
- Prior work applying computer vision within physical, hardware-based products.
- Solid grasp of video pipelines and how camera feeds behave in real deployments.
- Track record of taking vision models from development into live use, including object tracking across frames.
- Ability to own system architecture at a senior level.
- Self-directed working style with comfort collaborating across functions in a fast-moving startup.
- Bachelor's or master's degree in computer science.
Bonus points
- Exposure to industries where reliability and safety are critical.
Benefits
Compensation and benefits
- Base salary: USD 220,000 to 270,000 per year
- Equity
- Health insurance
- Retirement plan
- Commuter benefits
- Family planning benefits
Location and work model
- San Francisco, CA, United States
- Hybrid, 4 days in office
- Full-time