Alignerr

The AI training platform from the Labelbox team

Alignerr is the brand Labelbox uses to recruit contributors who evaluate and align the responses of AI models. It works like a job board: you sign up, apply to the projects that suit you, and depending on the role you either complete a technical assessment or upload a voice sample. It takes a wide range of profiles, from writing and translation through to scientific and technical specialities, and the hourly rates are high when there is work. The important caveat, worth knowing before you join, sits in exactly that "when there is work": availability comes in waves and projects are paused often.

How to get started step by step

1
Sign up and fill in your profile with real experience. Create your account at alignerr.com and complete your profile with your background and track record. You will see two brands side by side: Alignerr is the recruiting front and Labelbox is where the work actually happens, so the project dashboard lives in the Labelbox app. Have your ID ready, because verification includes a face recording.
2
Complete the assessments for what you actually know. There is no longer a general assessments tab: they are now only requested when a specific role requires one. Even so, the Alignerr team itself recommends completing every assessment that shows up on your dashboard in areas where you have prior experience, because those are what make you eligible for future projects.
3
Aim at specialist projects, not generalist ones. Generalist projects pile up enormous numbers of candidates, which is why the wait drags on. The ones demanding specific expertise (law, medicine, finance, engineering, particular languages) ask for more but are staffed by small teams, so your real odds are far better if your profile fits. Keep several applications open at once, because most never get resolved.
4
Expect an interview that digs into the why. The assessment does not stay on the surface: they ask why you made the decisions you made. Expect to walk through the data flow of a project on your CV justifying each technical choice, or to explain a concept as though teaching someone who just joined. Your reasoning matters more than being right: saying you do not know something and explaining how you would find out scores better than bluffing.
5
Check how the project bills before you start working. Look at whether the project pays per hour worked or per "finished hour", a model where only the time the system counts as productive is paid. Voice projects are where the widest gap between the advertised rate and the one actually received has been reported. Check too whether the initial training or evaluation phase is paid, because it is not always.
6
Treat it as a supplement and watch the dashboard. The unanimous advice from long-timers is not to rely on Alignerr as your main income. Some users strung together a year of steady work and then went months with nothing. Openings are announced through the dashboard and by email, so check both often, and keep another platform running alongside.

Pros

✓High hourly rates, far above surveys and microtasks
✓Backed by Labelbox, an established data-labelling company
✓Accepts a wide range of profiles, not only technical ones: writing, translation, and voice too
✓You can keep several applications open at once
✓Specialist projects face far less competition
✓Fully remote with no fixed schedule

Cons

✗Availability is very uneven: frequent pauses and dry spells lasting months
✗It usually takes weeks from signing up to a first project
✗Generalist projects are swamped with candidates
✗Under the "finished hour" model the real rate can land far below the advertised one
✗Initial evaluation phases are not always paid
✗Some users report accounts being closed without warning, in at least one case over the state or country they worked from
✗Identity verification requires a face recording

Tip to maximize earnings

The difference between waiting months and working with some regularity is almost always how specialised you are. If you hold a degree or specific experience, complete the assessments in that domain and apply only to those projects, because the generalist ones are overwhelmed. And accept the model before you join, which is what saves the frustration: this is contract work, not a job. Running it alongside micro1 is the sensible play, since when one of the two goes quiet the other usually has something open.

Get started on Alignerr