Nuro is looking to improve its machine learning infrastructure to support the development of its autonomous driving technology.
Requirements
- Strong coding skills in Python or C++
- Experience in building a cloud-based distributed training platform, which fully supports data and model parallelism.
- Experience investigating and optimizing training performance bottlenecks for deep learning models.
- Understanding of machine learning models and ML development life cycle
- Experience with Tensorflow, Keras, Pytorch and Cuda
Responsibilities
- Research and develop new distributed training frameworks and strategies to support training deep learning models with growing sizes.
- Improve model training speed by optimizing Tensorflow, Keras, Pytorch and Cuda kernel implementation.
- Design and implement world-class training frameworks for interactive, closed loop trajectory rollouts (Reinforcement Learning (RL) infrastructure)
- Engineer advanced tools to profile and monitor model training performance across all teams, to detect and triage training problems.
Other
- 2+ years of relevant work experience or an equivalent experience in PhD with 1+ years of relevant experience
- Annual performance bonus, equity, and a competitive benefits package
