Frontend Software Engineer

$40–50/hrrate as listed

micro1 recruits frontend engineers to evaluate and correct AI-generated code — reviewing model output in React, TypeScript, and modern CSS, then writing reference solutions and rationales that explain why one implementation beats another. Observed pay for this track is $40–50/hr, contract, remote and largely async.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

What the work involves

You work through tasks that put model output under an engineer's eye. A typical unit might be: a prompt asking for a virtualized list component, two candidate responses, and your job is to rank them, mark the specific defects (a stale closure in a `useEffect`, an accessibility attribute applied to the wrong node, a re-render cascade nobody measured), and write a correction that a senior reviewer would sign off on. Other task types include authoring original frontend prompts at a stated difficulty, repairing broken component code, and writing test cases that expose behavior a model got subtly wrong.

The hard part is rarely spotting that code fails — it's articulating why it fails in language a training pipeline can use. Vague feedback ("this is bad practice") gets rejected. Reviewers want the failure mode named, the browser or React behavior that causes it, and the fix stated concretely.

What the platform screens for

micro1 runs an AI-led interview before any human review. Expect it to probe:

Real production experience with React and TypeScript, not tutorial familiarity — you'll be asked to walk through decisions you personally made
Depth under follow-up: the interviewer pushes on rendering behavior, state management tradeoffs, bundling, and accessibility until you either demonstrate mechanism-level understanding or don't
Written clarity, since the deliverable is prose as much as code
Honest availability, stated in hours per week you can actually sustain

Logistics

Remote, contract, invoiced hourly against logged task time
Async — no standing meetings; queues open and close as projects ramp
Most contributors report 10–25 hours/week, with volume that fluctuates by project
Pay in the $40–50/hr range has been observed on this track; rates vary by project, assessed level, and specialization and are not guaranteed
Work is typically covered by NDA; you generally cannot publish or discuss task content

What they look for

Component Style Guides
Pixel-Perfect UI Implementation
Typescript
Responsive Web Design
CSS3

Before you apply

Would you pass this screen?

Drawn from what this listing actually requires. Answered in your browser — nothing is sent to us, and we do not store your answers.

Have you written and shipped React code in a production application, not in tutorials or side projects?
React in production · required
Have you written TypeScript code in a production codebase?
TypeScript experience · required
Can you commit to a minimum of 10 hours per week on a sustained basis?
Available hours per week · required
Are you able to write clear, technical English explanations of code behavior and defects?
Written English · preferred
Have you worked remotely on an async team or contract before?
Remote work experience · preferred

0 of 5 answered

Included with this listing

Interview prep — 11 questions

Matched to what this platform's screening actually measures. Answer each out loud in 60–90 seconds: claim → concrete example → implication.

What it tests: Whether you understand the actual costs of each option (re-renders, prop drilling, bundle size, cognitive load) and chose based on the problem, not cargo cult.
A strong answer: Name the specific tradeoff, state what each approach costs in concrete terms (e.g., 'moving to context would mean every filtered list re-renders when the filter changes'), and explain why you picked the one you did given your constraints.
Where people lose it: Describing only what you did, not what you didn't do or why — if you don't explain the road not taken, the interviewer can't tell if you chose deliberately or defaulted.