About this role
About the Role
Join the perception engineering team at an early-stage industrial automation startup, building computer vision systems for manufacturing quality inspection. You will work on real factory applications that identify parts and process steps, helping catch assembly errors before they lead to rework or defects.
What You'll Do
-
Build and improve computer vision pipelines for detection, classification, tracking, and temporal understanding.
-
Test first-person perception in challenging conditions such as camera motion, occlusion, motion blur, and changing lighting.
-
Use synthetic and real-world data to find failure cases and improve model performance.
-
Train, evaluate, and debug vision models in PyTorch.
-
Explore geometric vision, calibration, localization, pose estimation, and 3D perception.
-
Optimize models for low-latency inference and edge deployment, and build tools to evaluate real-world performance.
-
Support factory deployments and learn from system performance on the shop floor.
What We're Looking For
-
You are studying computer science, robotics, electrical engineering, mechanical engineering, or a related field, and are interested in computer vision and physical AI.
-
You are comfortable with Python and have hands-on experience in computer vision, machine learning, or robotics.
-
Experience with PyTorch, OpenCV, object detection, tracking, synthetic data, SLAM or VIO, 3D vision, CUDA or TensorRT, ROS2, or NVIDIA Jetson is helpful.
-
You have built and debugged technical projects and are ready to learn quickly and take ownership.
Compensation & Benefits
Visa sponsorship is available. Relocation from Germany is supported.
Location
On-site in San Francisco, California, United States.