micro1 is contracting visually trained specialists — photographers, art directors, designers, illustrators, curators — to rate and critique aesthetic images used to train generative models. The work is remote and asynchronous: you score batches of filtered images on composition, lighting, color, styling and originality, and write the reasoning behind each call.

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

What the work actually is

You receive batches of pre-filtered images and make calls on them: is this a genuinely strong photograph or a competently exposed one? Is the styling intentional or accidental? Is the composition original or a much-copied template? Most tasks pair a rating with a short written justification, and the justification is what the customer is buying — a model can already tell that two images differ, but it cannot say why one reads as art-directed and the other as stock. Some queues ask for head-to-head comparisons, others for single-image scoring against a rubric, and some for flagging images that are technically clean but aesthetically hollow (over-processed HDR, AI-generated tells, derivative trend-chasing).

Calibration is a real part of the job, not a formality. You will sit in async alignment rounds with project trainers and other visual specialists where disagreements get surfaced and the rubric gets revised. Guidelines change mid-project as the customer discovers what the model is over-rewarding, and people who treat a guideline update as an inconvenience rather than the point tend not to last.

What the screen looks for

micro1's process is AI-led: a conversational interview, sometimes a short visual judgment exercise, and a portfolio or track-record check. It is testing three things. First, whether your visual claims are anchored in real work — shot, art-directed, designed, curated, or juried, with specifics you can name. Second, whether you can articulate a judgment rather than assert it: "the light is beautiful" is worth nothing, "the key is a large soft source camera-left with a negative fill that keeps the jawline defined" is worth something. Third, whether you can hold a standard consistent across genres you personally dislike — a fashion editorial specialist who cannot fairly assess a still life is a calibration problem.

Logistics

Fully remote, contractor engagement, no fixed schedule; you work through batches on your own hours.
Observed pay $21–70/hr, not guaranteed — the band moves with the specific queue, complexity of the write-ups, and your calibration record.
Volume is project-driven and can be uneven; treat it as a supplementary engagement rather than a replacement for client work.
Expect an onboarding rubric, a paid or unpaid calibration round, and periodic re-alignment as guidelines evolve.

Visual Creative Expert

What the work actually is

What the screen looks for

Logistics

Interview prep — 10 questions