Working in AILesson 59 of 85

Contents

Working in AI

Lesson 59 of 85. Sign in to keep your place.

Module 6 · The interview, from first screen to signed offer

Pay, rates, and who actually decides them

8 min read

Read these numbers as shapes, not quotes

Salaries in this field vary by a factor of ten across countries and by a factor of three within a single city. Anyone quoting you one global number for “AI engineer salary” is selling something. What follows is rough, approximate for 2026, and should be checked locally before you act on it.

Junior analyst or junior data engineer, first role:

  • India: ₹4 to 9 lakh per year
  • Nigeria: ₦4 to 12 million per year
  • Philippines: ₱30,000 to 70,000 per month
  • Brazil: R$4,000 to 9,000 per month
  • Poland: PLN 8,000 to 15,000 per month
  • Germany: €45,000 to 60,000 per year
  • United States: $65,000 to 95,000 per year

Mid-level ML engineer, three to five years: roughly two to three times the junior figure in the same market. In the US, $130,000 to 200,000 is a common band.

Two distortions to keep in mind. First, the widely shared enormous packages come from a handful of frontier labs and big tech firms and describe perhaps a few thousand jobs worldwide. They are the tail, not the field. Second, employer type moves pay more than title does: in the same city, a bank, a startup, an outsourcing firm and a university will pay very different amounts for identical work.

Where the leverage is

Most people believe pay follows skill. It follows alternatives. The single largest lever in any negotiation is a second offer, and the second largest is not being the one who names a number first. “What is the band for this role?” is a complete and reasonable answer to “What are your expectations?”

If you are given equity in a startup, value it at zero when deciding whether the salary works. Most startup equity ends up worth nothing. Occasionally it is life-changing. Both facts are true and only the first should drive your rent.

Freelancing and contracting

Early rates on global platforms for data cleaning, automation, dashboards and scripting sit around $15 to 40 per hour, and experienced people with a clear niche charge $60 to 150. Local clients often pay less per hour but find you faster, pay sooner, and give you references in your own language.

A contract rate is not a salary. Do this arithmetic before comparing anything:

Text
$30/hour × 40 hours × 52 weeks     = $62,400   (fantasy)

Realistic billable weeks              38
Realistic billable hours per week     25   (selling, admin, gaps)
$30 × 25 × 38                      = $28,500
minus self-employment tax, own health cover,
unpaid sick days, unpaid holidays, one client
who never pays the last invoice

That is the honest comparison against a salary. Contracting can pay much better than employment, but the multiplier that makes it work is roughly two to three times the equivalent hourly salary, not the same number.

Protecting yourself

  • Take 30 to 50% before starting. A client who refuses is telling you something.
  • Bill in milestones, not one payment at the end.
  • Put the scope in writing, including what counts as “done” and how many revisions are included.
  • Do not hand over the final files, credentials or deployment before the final payment clears.
  • Cross-border work means currency swings and slow transfers. Price with a margin for both.

None of this is cynical. It is the ordinary equipment of getting paid, and every experienced freelancer learned it by losing money once.

Every lesson here is open. Your progress is recorded once you pass the AI at Work exam.

Check your understanding

Chidi contracts remotely from Lagos at $30 per hour. Lena is employed in Berlin at the equivalent of roughly $22 per salaried hour, doing the same work. Who is better off?

Pick the one you would defend. Your answer stays private.

Next lesson · 8 minWhat a rejection tells you, and what it does not

Stuck on something?Ask Rukmini, your AI tutor

How to cite: Addaly (2026) ‘Pay, rates, and who actually decides them’, in Working in AI. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/ai-careers/what-the-money-looks-like (Accessed: 31 October 2026).