$50–60/hr · Mercor · Hourly, 40 hours a week
You supply research materials and evaluation content to advance AI model training across multiple academic disciplines.
What you would do
- Gather technical information across STEM and non-STEM domains, then convert it into structured training materials
- Assess problem difficulty and calibrate questions appropriately for different learner skill levels
- Write clear explanations with minimal jargon to improve model learning efficiency
- Review your own work for logical consistency, factual accuracy, and instructional clarity
Who they want
- Background in 2-3 years of research, writing, or analytical work in any discipline
- Comfort with self-directed learning and ability to quickly master unfamiliar topics
- Strong written communication skills with attention to precision and nuance
- Current MacBook ownership with M-series chip or newer macOS 15+
- Availability for 10-20 hours per week of flexible, remote-friendly work
Main skills
What the interview asks about
1.Domain knowledge transfer
The interviewer assesses how quickly you absorb complex material from fields outside your expertise, since varied topics arise throughout the role.
For example: “You're assigned a physics question about quantum mechanics but lack formal training in the area - walk through how you'd research and then write an accurate explanation within 90 minutes.”
2.Quality calibration
Distinguishing between excellent and adequate work prevents poor training data from degrading model performance over time.
For example: “You're given three explanations for the same concept: one oversimplified, one precise and rigorous, one overly technical. Which would you submit to the training dataset and why?”
3.Self-directed workflow
Without daily check-ins, you must manage your time, catch errors, and maintain output quality without external accountability.
For example: “You notice a pattern of mistakes in your first ten submissions. How would you adjust your process to prevent the error from recurring in future batches?”
4.Technical writing fluency
Clear explanations that avoid jargon overhead are essential when material will train AI models used across diverse applications.
For example: “Rewrite this sentence for clarity: 'The epistemological foundations of synthetic reasoning modalities necessitate recalibration of categorical inference structures.' How would you know if your version works?”
A task you may get
Given a technical topic outside your background (e.g., immunology, blockchain, mechanical engineering), write three question-answer pairs at varying difficulty levels, rate each one, and justify your difficulty ratings.
How to prepare
- Review the published_skills list for your domain and refresh knowledge in two-three areas where you feel less confident
- Practice synthesizing information from three different online sources into a single coherent explanation under time pressure
- Prepare examples showing how you've learned complex material independently and applied it successfully in past work
The facts
- Pay
- $50–60/hr
- Hours
- Hourly, 40 hours a week
- Where
- Remote
- Field
- Data Analysis
- Posted
- 7/26/2026
- Places left
- 34365
We wrote this page from the public Mercor listing. It may be out of date, so read the full posting before you apply.