How I made $37k in 4 months on Mercor (And You Can Too)

my playbook for cracking the AI interviews, mastering RLVR, and becoming a top 10% earner.

There is a new category of jobs quietly emerging in the tech world. It doesn’t require you to live in Silicon Valley, and it doesn’t require a traditional corporate resume.

What it does require is deep domain expertise and the ability to teach artificial intelligence how to think.

Over the last four months, I’ve made $37,000 working on Mercor, a platform that connects experts with top-tier AI labs. I am currently a top performer at one of the leading AI training data companies and rank in the top 10% of earners.

This is my first article breaking down exactly how this ecosystem works—and more importantly, how you can break into it.

The Shift: The Demand for Human Intelligence is Real

Right now, Google, OpenAI, Anthropic, and others are locked in an arms race. But they’ve hit a bottleneck: they are running out of high-quality human data to train their models.

To bridge the gap from GPT-4 to the next frontier, these companies need experts for SFT (Supervised Fine-Tuning) and RLVR (Reinforcement Learning from Verifiable Rewards). They need humans who can evaluate complex agentic workflows and correct AI when it hallucinates.

The demand for human experts is massive, and the compensation reflects that.

The First Hurdle: Cracking the AI Interview

To get access to these high-paying tasks, you have to pass the vetting process. Here is where most people fail: Your interviewer is an AI.