General Motors is seeking to design and build scalable, reliable, and high-performance AI/ML platform infrastructure to support advanced AI research and model development initiatives for intelligent driving technologies across General Motors vehicles.
Requirements
- Strong programming skills in Python, with proficiency in frameworks such as,PyTorch (prefered), TensorFlow, or similar
- Experience with distributed computing, GPU computing, and cloud environments (AWS, GCP, Azure).
- 2+ years specialized experience in AI/ML infrastructure, e.g., enabling distributed training for scaling large ML models
- Extensive knowledge and experience with PyTorch 2.x+ and distributed training framework
- Experience with design and development of training framework that supports FSDP, Pipeline Parallelism and other scalable solutions to training large foundational models
- Experience with profiling, analysis, debugging and optimizing training and dataloading performance.
Responsibilities
- Design and development of scalable, reliabile, high-performance ML framework to support model training at scale.
- Model training performance analysis and optimizaiton solutions to scale distributed training workflows and maximize resource utilization across heterogeneous hardware environments, and save cost.
- Raise the bar on system observability, debuggability, and operational excellence, and user experience.
- Collaborate with cross-functional teams to integrate new features and technologies into the platform.
Other
- Bachelors or higher degree in Computer Science or equivalent major or equivalent experience
- 5+ years professional software engineering experience
- Willingness to travel to Sunnyvale, CA as needed
- Comfortable working in highly ambiguous and dynamic environments
- Excellent communication skills to resolve controversial, make consensus, communicate risks and give constructive feedback
