What the work actually involves

This is not an ongoing annotation role. It is a finite data-contribution task: you dig through your own camera roll, cloud backups and old phones to assemble 30 still selfies and 10 short video clips of your face, distributed across roughly the last five years so the set captures natural change over time — different hairstyles, weight, glasses, lighting, ageing. The critical constraint is authenticity: no beauty filters, no Snapchat or Instagram effects, no retouching, no AI enhancement. Where a filtered version is all you have, you are expected to locate the original file rather than submit the processed one. Expect a review pass, rejections on specific files, and a request to replace them before the payment triggers.

Because the output is biometric training data, the reviewers care about provenance and consistency. Files should come off a smartphone or camera with intact metadata where possible, show one clearly identifiable face (yours), and be sharp enough that facial landmarks are readable. Screenshots of photos, heavily compressed re-shares from messaging apps, and group shots cropped down to your face are the usual reasons a batch gets sent back.

What the platform screens for

The screen is short and mostly factual. It confirms that you genuinely have a multi-year archive rather than plans to shoot everything today, that you understand and consent to how facial data will be used, that you can follow a written spec without improvising, and that you can be reached to fix rejected files. Micro1's referral angle means you may also be asked whether you have siblings or twins who would contribute — identical twin pairs are unusually valuable to this kind of dataset, and saying so honestly is worth more than overstating it.

Logistics

  • Fully remote and fully asynchronous; no calls, no shifts.
  • Realistic effort: two to four hours of searching, sorting and uploading, plus a smaller follow-up pass on rejections.
  • Compensation is a single flat payment on acceptance of the complete set. The amount is not disclosed in the listing and rates for facial-data collection vary widely by platform and cohort — treat any figure you see quoted elsewhere as observed, not promised.
  • Read the consent terms before you upload. Facial data submissions are generally not retractable once ingested, and that is the part of this quest that deserves the most thought.