Real-Time Computer Vision Architect

What you'll do

• Break down high-level product goals into system, module, and component requirements including firmware, ISP, and OS constraints.
• Develop and optimize computer vision perception algorithms and ISP pipelines for performance, power, and memory on embedded platforms.
• Evaluate sensing tradeoffs across cameras, depth sensors, and IMUs with firmware, mechanical, electrical, and optical science teams.
• Debug and optimize performance at the firmware or OS level to match hardware capability with algorithmic needs.
• Tune ISP behavior for tasks like noise reduction, color correction, demosaicing, and dynamic range enhancement.
• Translate ambiguous product goals into detailed feature-level architectures and defensible end-to-end system requirements.
• Integrate modern techniques like vision transformers and foundation models into resource-constrained environments.

What we're looking for

• BS in Computer Science, Electrical Engineering, or other relevant fields.
• Minimum of 10 years of relevant, proven experience.
• Proven track record in developing and improving CV perception algorithms and ISP pipelines.
• Demonstrated experience optimizing computer vision algorithms for performance, power, and memory on embedded platforms.
• Strong understanding of perception sensing technologies including color and monochrome cameras, depth sensors, and IMUs.
• Depth of expertise in one or more computer vision specialty areas such as SLAM, tracking, classification, or scene understanding.
• Experience in translating ambiguous product goals into detailed feature-level architectures and system requirements (preferred).
• Exposure to vision transformers, foundation models, and vision-language models in resource-constrained environments (preferred).