Yes. Health AI training and evaluation work is carried out remotely, in your own time, at times of your choosing.

Do I need technical or coding skills?

No. Clinical judgement and real-world experience are what is required. Most tasks involve reading, reasoning, and writing assessing AI outputs, defining reference answers, and quality checking responses.

Is patient data involved?

No identifiable patient data is used.

Is this a job or freelance contract work?

This is contract project work, not employment. Most arrangements are flexible and task-based rather than salaried. You will be contracting directly with Mercor or Micro1, not with this site.

Can I do this alongside NHS or locum work?

In many cases yes subject to your employer's policy on secondary work and any conflicts of interest.

Your CV for AI Training Work: Set Out Your Entire Stall

By Sean Key, Editor — Applied Clinical Judgement. Last reviewed July 2026.

Here is the mindset that gets people the most work, and it is not how anyone is taught to write a CV. You never know what a project is going to ask for. One month the work needs financial reasoning, the next it needs someone who understands cricket, or home cooking, or tabletop games, or caring for an elderly relative. So the job of this CV is not to present a tidy career. It is to set out your entire stall – everything you genuinely know, in depth – so that whatever comes up, you are already a match.

That means this CV does two jobs most people only prepare it for one of. It has to be read by software that pulls out your skills and experience to match you to work. And it has to stand up in an interview built directly from what it says, because the AI interviewer will ask you to defend your own claims. A CV that reads well but cannot survive questioning will cost you at the second stage.

This page covers all of it: how to make your CV machine-readable so you get matched, how to make it specific enough to defend, and – the part people get wrong – how to make it broad enough that you match to far more than your job title alone.

Make it machine-readable first

When you upload your CV to Mercor or Micro1, the platform parses it to pre-fill your profile – pulling out your roles, skills, and education. If the parser cannot read it cleanly, your profile starts out thin, and a thin profile gets matched to less work. So the format matters before the content does.

Use a searchable PDF, not a scan or an image. The text must be selectable. A photographed or image-based CV gives the parser nothing to read.
Keep it to one or two pages. Long CVs dilute the signal and parse less reliably.
Use plain, clear headings – Experience, Skills, Education, Certifications. Avoid tables, columns, text boxes, and graphics for the core content; parsers often mangle them.
Name your tools, titles, and dates explicitly. Exact role titles, employment dates, and the specific technologies or methods you used give the matching system something concrete to work with. “Analyst” tells it little; “Financial analyst, equity research, 2021-2024, Bloomberg and Python” tells it a lot.

After you upload, review what the platform imported. Both let you correct the parsed result, and their own guidance is to add anything it missed – especially outcomes, metrics, and your skill set. Do not skip that step; the pre-fill is a starting point, not the finished profile.

Then make it specific enough to defend

Here is the part that trips people up. The interview is generated from your CV, and it follows up on what you claim. If your CV says you “led a project”, expect to be asked which project, what your role actually was, and what happened. Vague lines invite questions you cannot answer, and that hurts your score.

So write lines you can stand behind:

Lead with quantified outcomes. “Improved X by Y% over Z months by doing A” beats “responsible for improving performance”. Numbers give you something concrete to talk about and signal real ownership.
Own your actual part. Claim what you did, not what your team did. The follow-up will find the gap between “we” and “I”.
Be specific about scope. Name the domain, the scale, the tools. Specificity is what makes both the parser and the interviewer treat your experience as real.
Cut anything you cannot discuss in depth. If you would not be comfortable being questioned on a line, take it out. A shorter CV you can defend fully beats a padded one with soft spots.

A note on how it reads, not just what it says. These platforms assess clear, concise writing from the very first thing they see, and your CV is that first thing. Tight, well-organised lines that make their point and stop are themselves a signal you can do the work, which is largely about writing clearly and following a brief. A rambling, padded CV works against you twice: it parses worse, and it suggests the opposite of the skill they are hiring for. Say what you did plainly, and trust the specifics to do the work.

Match your CV to the work

Where a platform shows the language of the roles or skills it wants, echo that language honestly where it genuinely applies to you. The matching systems look for alignment between your profile and the listing, so using the same terms for the same things – when they are true of you – helps you surface for the right work. This is not keyword-stuffing; it is describing your real experience in the words the system is matching on.

Put more on this CV than you would on LinkedIn

Here is a difference worth understanding. Your LinkedIn is a public professional shopfront, pointed at your career. This CV is not that. It is training data about everything you know, and the range of work these platforms pay for is far wider than your job title. So things you would leave off a career-focused LinkedIn belong here.

Two kinds of thing in particular:

Earlier or “off-track” experience, especially if you changed career. If you retrained, the work you did before is not irrelevant here – it is a second domain of real expertise you can be matched on. A nurse who used to be an accountant, an engineer who used to teach, a lawyer who spent years in retail: put all of it in. Each past field is another set of projects you could qualify for. On LinkedIn you curate for one direction; here you are declaring range.
Everyday skills, hobbies, and interests. This surprises people, but these are live task categories. AI training projects come and go across an enormous spread of ordinary human knowledge, and someone with genuine depth in a topic is exactly who they want. Over time these have included things like household budgeting, cooking and recipes, holiday and event planning, caring, DIY and home repair, and housework method – alongside a long list of pastimes: tabletop role-playing games and Dungeons and Dragons, cricket, American football, soccer, amateur dramatics, singing and playing an instrument, podcasting, travel, and more. If you have real, describable depth in something, it is worth a line.

The honesty rule still holds: only list what you genuinely know and could talk about, because the interview and the tasks will test it. The test is simple – is it something you are genuinely passionate about or have real in-depth knowledge of? If yes, it earns a line, however unprofessional it might feel on a normal CV. If it is a passing interest you could not be questioned on, leave it. This is not an invitation to pad. It is a reminder that “qualified” here is far broader than a CV usually assumes, and leaving out a real area of knowledge just means you match to less work.

Why this breadth pays

The reason setting out your whole stall works is that a lot of this work is generalist by design. Alongside the specialist projects, there is a steady stream of general evaluation work – judging whether an AI’s answer is good, accurate, and sensible across everyday topics – and that work draws on exactly the wide, ordinary knowledge most people leave off their CV. The broader and more specific your stall, the more of that general work you match to, on top of your specialist areas. There is more on the evaluation skills these roles use, and how to build them, in our guide on assessments and training.

One CV, one story

Breadth is not the same as inconsistency, and it is worth being clear on the difference. Adding your old career and your genuine interests widens the range of work you match to – that is good. What counts against you is contradiction: claiming a level of expertise your interview answers cannot back up. On Micro1 in particular, the system cross-references your CV, your LinkedIn, and your spoken answers, and if your CV claims deep specialism in something your examples cannot support, that shows up as a consistency problem. So list widely, but claim honestly – every line should be something you could talk about if asked. A broad CV of things you genuinely know is a strength; a CV that overstates any one of them is the risk.

Frequently asked questions

Does the CV format really matter, or just the content?

Both, and format comes first. If the platform cannot parse your CV cleanly, your profile starts thin regardless of how good the content is. Use a searchable PDF, one or two pages, plain headings, no graphics in the core content.

How long should my CV be?

One or two pages. Longer CVs parse less reliably and dilute the signals the matching system uses.

Should I tailor my CV to each role?

Where you can, echo the real language of the role for things that genuinely apply to you – it helps the matching system align your profile with the listing. Do not invent alignment you do not have; the interview will test it.

What is the most common CV mistake on these platforms?

Vague, unquantified claims you cannot defend. The interview is built from your CV and probes it, so a line like “led key initiatives” with no specifics invites questions that expose the gap. Be concrete or cut it.

Should I include hobbies, interests, or my old career?

Yes – much more than you would on a normal CV or LinkedIn. This CV is training data about everything you know, and projects span a huge range of ordinary knowledge, so a former career, a serious hobby, or an everyday skill can each match you to work. List anything you genuinely know well enough to be questioned on. Leaving real areas of knowledge off just narrows the work you can be matched to.

Will putting hobbies on my CV look unprofessional?

Not here. These platforms are not a traditional employer judging your polish – they are matching what you know to tasks that need it. A CV that reads as broad and specific is an advantage, provided every line is something you could actually discuss.

Do not polish it forever

One practical warning, because this is where people lose the most time. A CV is never truly finished, and it is easy to spend a fortnight nudging wording while your application sits unsubmitted. Do not. Get it to a solid, honest, specific standard – parseable, quantified, broad, defensible – and then submit it. Aim to have your first application in within two or three days of starting, not two or three weeks. On platforms that fill their qualified places in the order people apply, a good CV submitted this week beats a perfect one submitted next month. You can keep improving it afterwards, and should – but a live application is working for you, while a draft you are still tweaking is doing nothing.

The bottom line

Treat your CV as two things at once: a machine-readable profile that gets you matched, and a script you will be asked to defend. Make it parseable, make every line specific and true, and cast it wider than a normal CV – your old career, your real hobbies, the everyday things you genuinely know are all matchable here. Just make sure every line is something you could talk about if asked. Do that and you clear the first two gates in one move.

This is one part of our complete guide to getting hired for AI training work, which covers your CV, LinkedIn, the interview, getting matched, and maximising your work.

—

Applied Clinical Judgement is a referral service run by Sean Key, a Digital Health Senior Programme Manager with 29 years of NHS and private-sector experience. [LinkedIn] · [Book a vouching call] · [Application guide (PDF)]. ACJ is not an employer or recruiter; it points you to the platforms that do the hiring, and may receive a referral credit at no cost to you.