Graph Compiler Engineer

OpenAI is seeking a Graph Compiler Engineer to join their Platform ML team. This role focuses on building the ML side of OpenAI's state-of-the-art internal training framework used to train cutting-edge models. The successful candidate will work on distributed model execution, interfaces, and implementation for model code, training, and inference.

Key Responsibilities:

• Work on OpenAI's internal training framework to enable easy graph capture of research experiments
• Apply advanced optimizations to the graph to maximize hardware resource utilization
• Collaborate with researchers at the cutting edge of ML research
• Improve training throughput and researcher productivity

Ideal Candidate:

• Strong software engineering skills with proficiency in Python
• Experience running machine learning experiments
• Experience working on a machine learning compiler
• Deep knowledge of GPU performance and systems-level optimization

The role offers an exciting opportunity to work on OpenAI's training framework and contribute to accelerating progress towards AGI. Small improvements in hardware utilization for large training runs can produce significant cost savings, while enabling researchers to iterate quickly on their ideas will lead to substantial improvements in model capabilities and compute efficiency.

OpenAI offers a competitive compensation package, including a salary range of $245K – $385K, generous equity, and comprehensive benefits. The company is committed to diversity, equality, and creating an inclusive work environment.

This is a hybrid position based in San Francisco, CA, requiring 3 days in the office per week. Relocation assistance is available for new employees.