Cavnue’s mission is to unlock autonomous mobility and save lives, time, and money. We do this by building the roads of the future.
As the Computer Vision Lead at Cavnue, you’ll bring critical skills to a team working to decode events on the road through distributed sensors in the built environment. We’re looking for people who have spent time on camera calibration, setting up data streams, and dealing with image stabilization in real world environments under less-than-ideal conditions. Help us build and solve some of the most complex, near real-time coordination problems at scales that matter to everyday people using our streets.
- Working closely with the hardware engineering team to solve problems in camera systems in dynamic, real world environments
- Developing applied solutions for real-world, complex problems in autonomous robotics and road transportation
- Driving efforts on camera calibration, image segmentation, camera based scene understanding, and mapping camera data to a global and local frame
- Creating the framework for decomposing issues with imagery at every stage of the acquisition, inference, compression, transfer, retention, sampling, training, and storage processes.
- Working with the Product and Engineering teams to ensure that the right data sets are being collected for relevant tasks at varying geospatial and temporal scales
- Participating in the engineering lifecycle and designing high-quality ML infrastructure and data pipelines, defining production code standards, conducting code reviews, and working alongside infrastructure, reliability, and hardware engineering teams.
- Expanding Cavnue’s competitive advantage through understanding, applying, or inventing novel ML and CS techniques
- MS or Ph.D. in Computer Science, Electrical Engineering, Statistical Signal Processing (or related field)
- 7+ years of work experience with imagery in the autonomous vehicles, biomedical, remote sensing, or robotics industry
- Track record of solving complex problems in computer vision in real-world systems, including embedded systems
- Deep understanding on geometry-based Computer Vision approaches (Structure-from Motion, Stereo vision, SLAM, Visual Odometry)
- Understanding of camera models, calibration methods, distortion models and rectification methods.
- Detection and classification experience relevant to the driving task (objects, lanes, signs, etc.)
- Embedded software development and optimization experience
- Experience in deployment of real-time applications
- Statistical signal processing and sensor fusion experience
- Firm grasp of large-scale data structures and data pipelines, data modeling, software architectures, and the latest ML libraries and frameworks e.g. TensorFlow, Pytorch.
- Experience with low-level libraries for image analysis (OpenCV), scientific programming (NumPy), and non-linear solvers (e.g., Ceres)
- C++ and Python experience required. Jupyter experience ideal
We are building an incredible team of employees with diverse backgrounds and experiences. We believe that great work can occur anywhere and are open to considering candidates who meet our needs who reside outside our geographic footprint. We also value the impact that can result from co-located teams and some roles may require regular presence in one of our offices.
Cavnue is an Equal Opportunity Employer and prohibits discrimination or harassment of any kind. All employment decisions at Cavnue are based on business needs, job requirements, and individual qualifications, without regard to race, color, national origin, sex, gender, age, religion or belief, disability, sexual orientation, family or parental status, veteran status, or any other status protected by law.