General MotorsRemote
Senior Machine Learning Engineer - ML Training Infrastructure
United States of America
Full Time
4 hours ago
$170,000 - $240,000 USD
H1B Sponsor Likely
Key skills
Machine Learning Training InfrastructureDistributed TrainingPythonPyTorchTensorFlowGPU ComputingCloud Environments - AWSCloud Environments - GCPCloud Environments - AzurePerformance Profiling and OptimizationPipeline ParallelismFully Sharded Data Parallel (FSDP)AIMachine LearningMLAWSAzureGCPCommunication
About this role
General Motors is seeking an experienced Senior Machine Learning Engineer specializing in ML Training Infrastructure. This role involves designing and building scalable AI/ML platform infrastructure to support advanced AI research and model development initiatives.
Responsibilities:
- Design and development of scalable, reliable, high-performance ML framework to support model training at scale
- Model training performance analysis and optimization 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
Requirements:
- Bachelors degree or higher in Computer Science or equivalent major OR equivalent relevant experience
- 3+ years professional software engineering experience
- 2+ years specialized experience in AI/ML infrastructure, e.g., enabling distributed training for scaling large ML models
- Strong programming skills in Python, with proficiency in frameworks such as, PyTorch (preferred), TensorFlow, or similar
- Experience with distributed computing, GPU computing, and cloud environments (AWS, GCP, Azure)
- Willingness to travel to Sunnyvale, CA as needed
- Comfortable working in highly ambiguous and dynamic environments
- 3+ years of professional software engineering experience
- Self-motivated, strong execution, impact-delivering oriented
- 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 data loading performance
- Excellent communication skills to resolve controversial, make consensus, communicate risks and give constructive feedback