Good, Next
Career guides

How to get an AI job without 5 years of experience

6 min readUpdated 19 August 2026

Job postings in AI and data roles are notorious for listing experience requirements that don't match the reality of the field. A lot of the tooling and best practices are only a few years old, so '5+ years of production ML experience' is often written by someone reusing an old template, not a hard requirement enforced at the application stage.

Read the requirement as a wishlist, not a gate

If you meet most of the concrete requirements (the languages, the type of problem, the general seniority) apply anyway. Hiring managers routinely interview candidates who don't hit every listed year count, especially at companies that are growing fast enough to need people now rather than in six months.

Build evidence that substitutes for tenure

  • A small number of real projects, finished and documented, beats a long list of tutorials followed.
  • Contributing to an open-source project in the space you want to work in is a visible, checkable signal.
  • Writing publicly about something you built, including what didn't work and not just the highlight reel, reads as more credible than a polished portfolio alone.
  • Freelance or contract work, even unpaid or for a friend's project, counts as real experience if you can describe the problem and the outcome clearly.

Target the right kind of company

Early-stage startups and scale-ups are generally more open to strong signal over exact years, because they're optimizing for someone who can move fast, not someone who fits a rigid ladder. Large enterprises and heavily regulated companies tend to be stricter about formal experience. Not always, but often enough to weight your effort accordingly.

In the interview, own the gap directly

Don't pretend the experience gap doesn't exist. Address it in one sentence and move straight to evidence: 'I haven't shipped this at the scale your job description mentions, but here's a project where I dealt with a very similar constraint.' Confidence about what you don't know yet is more convincing than trying to talk around it.

Stop manually deciding what to apply to

Tell Good, Next what you're looking for and get your next three actions.