Software Engineer I ( Research Engineer )

Meta Platforms • Menlo Park, CA

6d • Remote

About The Position

We're looking for a software engineer to help build and evaluate CUA (Computer Use Agents) — models trained to operate real software and complete computer-based tasks the way a skilled engineer would. You'll own complex, long-running technical workflows end to end — implementation, validation, documentation, and follow-through — and design and maintain the tasks, benchmarks, and test suites that measure CUA capability as the underlying models, APIs, and infrastructure keep evolving. You hold an exacting bar for code quality: you love writing tests, you're fluent with deployment strategies like canary releases and rollbacks, and you bring a strong sense of risk management. You have a sharp eye for weak implementations and give (and take) blunt, direct feedback. You're also fluent with AI coding assistants like Claude Code, Codex, or Cursor, with your own well-developed best practices for using them, and you're proficient in Rust, Python, and/or TypeScript. Above all, you bring strong debugging and experimental discipline — the ability to investigate discrepancies across code, configuration, infrastructure, and results within the CUA stack, and turn ambiguous findings into reproducible conclusions.

Requirements

• Passionate about coding since childhood with a drive to dive deep into technology.
• Maintain extremely high standards for code quality.
• Love writing tests.
• Familiar with deployment strategies like canary releases and rollbacks.
• Possess a strong sense of risk management.
• Familiar with at least one AI coding assistant (e.g., Claude Code, Codex or Cursor) and have developed your own insights and best practices for using them.
• Proficient in one or more of the following languages: Rust, Python, or TypeScript.
• Possess a sharp eye for spotting "garbage" code/implementation.
• Willing to give blunt, direct feedback (call out bad code) and are equally thick-skinned enough to receive it.
• Strong debugging and experimental discipline.
• Comfortable owning complex, long-running technical workflows end to end, including implementation, validation, documentation, and follow-through.

Nice To Haves

• Experience in server-side or infrastructure development, with a track record of building highly available and stable systems.
• Exceptional communication skills.
• Able to engage effectively with algorithm teams, product managers, operations, and executives.
• Can translate complex technical concepts into language that absolutely anyone (technical or non-technical) can easily understand.
• Have your own carefully maintained open-source project(s)—the number of GitHub stars doesn't matter.
• Experience with ML evaluation, distributed systems, developer infrastructure, or large-scale testing.
• Experience maintaining benchmarks or test suites while underlying models, APIs, and infrastructure evolve.

Responsibilities

• Own complex, long-running technical workflows end to end — implementation, validation, documentation, and follow-through.
• Design and maintain the tasks, benchmarks, and test suites that measure CUA capability as the underlying models, APIs, and infrastructure keep evolving.
• Investigate discrepancies across code, configuration, infrastructure, and results within the CUA stack, and turn ambiguous findings into reproducible conclusions.

What This Job Offers

• Job Type: Full-time
• Career Level: Entry Level
• Education Level: No Education Listed