Software Engineer, ML Systems & Training Architecture

OpenAI•San Francisco, CA

8d•$295,000 - $380,000•Onsite

About The Position

The OpenAI Robotics team is focused on unlocking general-purpose robotics and pushing towards AGI-level intelligence in dynamic, real-world settings. Working across the entire model stack, we integrate cutting-edge hardware and software to explore a broad range of robotic form factors. We strive to seamlessly blend high-level AI capabilities with the constraints of physical systems to improve peoples’ lives. As a Senior Software Engineer, ML Systems & Training Infrastructure, you will be a deeply hands-on engineering force multiplier for the robotics team. You will help keep the training framework and surrounding infrastructure healthy, review and improve code quickly, debug failures across ML systems and infrastructure, and unblock researchers and engineers when the path from idea to working training job gets rough. We’re looking for people who love writing, reading, reviewing, and fixing code; who can get productive quickly in unfamiliar systems; and who bring strong practical judgment without a lot of ego or process overhead. This role will be based in San Francisco, CA and be expected in office 5 days per week and offer relocation assistance to new employees.

Requirements

Strong software engineering fundamentals and excellent code review judgment.
Experience with ML systems, training frameworks, GPUs, distributed systems, infrastructure, or similarly complex technical environments.
Read and debug unfamiliar codebases quickly, and enjoy getting to root cause.
Ship high-quality code with strong velocity and pragmatic judgment.
Low-ego, responsive, and motivated by helping researchers and engineers move faster.
Prefer being a highly effective hands-on IC over driving broad process-heavy initiatives.
Experience reviewing messy, fast-moving, or AI-generated codebases.

Responsibilities

Review, improve, and clean up code across training frameworks and adjacent infrastructure.
Identify risky or low-quality changes before they land, and raise the code quality bar without slowing the team down.
Debug issues across ML training systems, GPUs, clusters, networking, and related infrastructure.
Help researchers and engineers unblock broken training jobs, flaky workflows, and brittle internal tooling.
Improve the reliability, maintainability, and usability of the robotics team’s training framework.
Move quickly on practical engineering problems that directly affect team velocity.