Senior Research Infrastructure Engineer

present of AI; San Francisco (CA); Hybrid; USD 180,000 - 240,000; Full time

Job summary

HUD is seeking a Senior Software Engineer to build tooling for AI experiments. You will create reliable environments, manage trajectories, and scale evaluation while turning research code into reusable software.

Strong Python and backend skills, plus experience with Docker, Linux, and cloud platforms, are essential. You will partner with researchers to translate vague needs into practical, maintainable solutions that accelerate AI research.

Qualifications

• Strong Python and backend engineering experience.
• Comfort with Linux environments and cloud systems.
• Experience collaborating with researchers to clarify needs.

Responsibilities

• Build tools and systems enabling researchers to run AI experiments reliably.
• Ensure environment reproducibility, trajectory collection, and scalable evaluation.
• Turn rough research code into tested, reusable software assets.

Skills

Python

Backend development

Linux

Cloud platforms

Tools

Docker

Job description

AI & Data · San Francisco / Singapore / Remote (Asia) / Remote (North America) · Full-time · Senior level

This role is about building the tools and systems that let researchers run experiments on AI agents without fighting their own setup. That means reliable environment runs, trajectory collection, evaluation at scale, and turning rough research code into tested software others can reuse. A strong candidate brings solid Python and backend skills, comfort with Docker, Linux and cloud systems, and the habit of working closely with researchers to shape vague needs into useful products.

How it fits the AI era

AI agents are now trained and judged through large numbers of simulated runs, so the quality of the infrastructure behind those runs shapes how fast research moves. AI coding tools can speed up writing pipelines and tooling, but deciding what to build, spotting why a run failed, and making results trustworthy still takes human judgment. Understanding what researchers are really asking is a human skill that this role leans on heavily.

How it builds AI experience

The work sits close to how frontier AI agents are trained and tested, since the company builds and runs reinforcement learning environments. Someone in this role would see agent failure modes, data quality problems and evaluation methods up close. To build this experience beforehand, try running a small reinforcement learning environment, collecting rollouts, and setting up a repeatable evaluation of an agent.

Worth showing

• Maintainable Python systems


• Backend or pipeline ownership


• Docker and cloud debugging


• Tools that sped up researchers

This is our own read, written from the employer's posting and the company's Spotlight. The employer's posting is the source of truth for the role.

About HUD

A platform for building, running and selling the reinforcement learning environments that teams use to train and test AI agents.

• Design Engineer · San Francisco / Singapore / Remote (Asia) / Remote (North America)


• Technical Writer · San Francisco / Singapore / Remote (Asia) / Remote (North America)


• GTM Engineer · San Francisco / Singapore / Remote (Asia) / Remote (North America)