On our Perception team, you have the opportunity to work with world-class ML engineers and research scientists, whose mission is to make self-driving vehicles a reality and to create a positive social impact. Our team works on the tech stack responsible for perceiving the dynamic scenarios, and further tracking and classifying objects around our robo-taxi. We are looking for engineers who are passionate about Level 4 autonomous driving technology, excited by intellectual challenges, and interested in pursuing career growth with a fast-growing company.
What you'll be doing:
- Productionize and deploy our perception models into edge devices
- Develop and optimize GPU-accelerated algorithms using CUDA
- Develop advanced post-processing and classical vision algorithms
- Drive performance optimization at system level (latency, throughput, memory, determinism)
- Drive code quality through reviews, testing strategies, and performance benchmarks
- Collaborate deeply with ML engineers to:Improve model robustness and deployment readinessInfluence training objectives based on downstream system needsDebug and resolve model-system integration issues
What We're Looking for:
- Bachelor's Degree. Preferably Masters or Ph.D. in Machine Learning, Computer Science, Robotics, Applied Mathematics, Statistics, Physics or a related field or equivalent industry experience
- Strong proficiency in modern C++ (C++14/17 or later)
- Hands-on experience with CUDA and GPU performance optimization
- Deep understanding of modern and classical computer vision and image processing techniques
- Experience with performance profiling and debugging tools (e.g., Nsight, Valgrind, perf)
- Proven track record of developing and deploying perception systems for autonomous vehicles or robotics
- Fluency in Python, including standard scientific computing libraries and Python bindings development experience
- Advanced knowledge of software engineering principles including software design, source control management, build processes, code reviews, testing methods
