What the work actually involves

You work against real open source repositories rather than toy prompts. A typical task is scoped like an issue on a tracker: implement a feature in an existing codebase, diagnose and fix a non-obvious bug, or refactor a module for performance and readability without changing behaviour. Languages in rotation include Python, Java, Rust, C++, TypeScript, Go and C#; most contributors are staffed against the two or three they know deepest. Alongside the diff, you write a technical explanation of the design — why this approach, what alternatives you rejected, what the failure modes are. That write-up is often worth as much to the customer as the code, because it is what teaches a model to reason rather than pattern-match.

Some tasks are validation rather than authoring: reviewing a proposed change for correctness, output consistency, and adherence to project conventions, and saying precisely where it goes wrong. Competitive-programming instincts help here — recognising complexity traps, edge cases, and off-by-one boundaries quickly.

What the screen looks for

  • Verifiable open source history. Named repositories, merged PRs, issues you triaged. Vague claims of "contributing to OSS" do not survive follow-up.
  • Depth in at least two languages, including their idioms, tooling, and memory or concurrency models — not just syntax.
  • Debugging narrative. Expect to be walked through a real bug you fixed, from symptom to root cause to test.
  • Written clarity. The screen samples your explanation quality directly; terse answers read as a risk for the write-up portion of each task.

Logistics

Fully remote, asynchronous, contractor engagement. Compensation is per accepted task, not per hour, so effective hourly rate depends on your speed and on how often your submissions clear review on the first pass. A weekly minimum number of submissions applies. micro1 moves fast on hiring — roles are typically filled within 48 hours, and selected experts are expected to begin tasks within 24–48 hours of finishing onboarding. No prior AI or ML experience is expected or required.