Overview:
As a Senior ML Engineer, you will design and build scalable, high-performance ML training infrastructure to enable AI research and model development for intelligent driving. You will work with ML engineers, researchers, and cross-functional partners to raise observability, performance, and reliability of the platform. You’ll optimize distributed training across heterogeneous hardware and cloud environments, driving cost efficiency and research impact. This role offers hands-on technical ownership and collaboration on state-of-the-art AI solutions shaping GM’s autonomous driving initiatives. You will operate in a dynamic, cross-functional setting with a strong emphasis on execution and impact Compensation / Benefits health and wellbeing benefit programs medical, dental, vision retirement savings plan paid vacation & holidays tuition assistance programs GM vehicle discounts
Responsibilities:
Design and develop scalable, high-performance ML training framework for model training at scale Improve observability, debuggability, and operational excellence of the platform Analyze and optimize training performance to maximize resource use and reduce cost Collaborate with cross-functional teams to integrate new features and technologies Support and improve the training infrastructure for AI research and model development initiatives
Key requirements:
Bachelor in Computer Science or equivalent, or equivalent relevant experience 5+ years software engineering experience 1+ years of AI/ML infrastructure experience, enabling distributed training for large models Strong Python programming skills with PyTorch (preferred) and/or TensorFlow Experience with distributed computing, GPU computing, and cloud environments (AWS, GCP, Azure) Willingness to travel to Sunnyvale, CA as needed Comfort in ambiguous, dynamic environments Self-motivated, execution-driven Excellent communication and ability to resolve conflicts and give constructive feedback Collaborative, cross-functional teamwork Python programming PyTorch 2.x+ and distributed training frameworks TensorFlow or similar