What the work involves
Day to day, you are reading frontend code you did not write and deciding whether it is correct, idiomatic, and safe to ship. That means comparing two model responses to a component spec and justifying which is better; catching the React bug that only shows up on re-render or in a stale closure; flagging accessibility and layout failures a passing test suite happily ignores. On authoring projects you go the other direction — writing prompts, component specs, and reference implementations with tests, so there is a ground truth to score against.
The grading is rubric-driven and the written justification matters as much as the verdict. A one-line "Response A is better" is not usable training signal. Reviewers are expected to name the failure precisely — unnecessary `useEffect`, `key` on array index, unhandled loading state, `any` smuggled into a public prop type — and to be consistent across dozens of samples in a sitting.
What the screen looks for
micro1's intake is conducted largely by an AI interviewer that asks about your background and then follows up on specifics. It is testing whether your frontend experience is real and recent: which React version and rendering model you actually shipped on, how you handled state and data fetching, what you would change about a decision you made. Vague seniority claims get probed rather than accepted. Expect at least one code-judgment exercise where you critique a snippet, and questions about throughput and availability, since project work arrives in batches with deadlines.
Logistics
- Fully remote and largely asynchronous; batches often have a turnaround window rather than fixed shifts.
- Typical engagements run 10–25 hours per week, sometimes more when a project ramps.
- Rates are observed in the $90–140/hr band and vary by project and task complexity; nothing here is guaranteed.
- You will need your own machine, a reliable connection, and comfort working in browser-based review tooling.