JSON-fluent & Rubrics Expert (Train AI Models Part Time!)
hackajob · United States
- Hybrid
- Contract
- $40,000 / year
- United States
Tailored resume — keyword-matched to this role.
Hiring manager — we find who's hiring.
Intro email — drafted to reach them directly.
Job highlights
- Assess AI model outputs using JSON and rubrics.
- Requires immediate start for a weekend sprint.
- Must be US-based and detail-oriented.
- Dedicate 20-40 hours weekly for 12-24 hour task turnaround.
- Must use a desktop or laptop computer.
About the role
About the Opportunity
Mercor is recruiting detail-oriented graders for a weekend sprint beginning today, July 24th. Please apply only if you can begin work immediately! Onboarding happens right away, and the sprint starts on day one. This project is supporting one of the world’s leading AI research organizations. You will assess model outputs against a defined rubric of grading criteria. The ideal candidate has experience processing JSON files, and an understanding of assessing and grading human data tasks against rubrics.What you’ll do:
You will assess JSON-formatted AI model outputs through the following:- Reading and understanding written task descriptions
- Reading and understanding written task rubrics for these tasks
- Parsing, reading, and understanding a JSON-based output, and grading its contents against the task rubrics provided
Applicants must:
- Be available to start today, July 24, running to July 27-28, with the possibility for extension later into the week.
- Be comfortable reading and interpreting JSON files.
- Ideal candidates will have prior Mercor experience, especially in applying detailed rubrics and evaluation criteria consistently.
- Be based in the US.
More details about the role:
- You must be able to dedicate a minimum of 20-40 hours per week on this task.
- Turnaround expectation: Ability to complete tasks within 12-24 hours.
- You will need access to a desktop or laptop computer for this project.
- We cannot accept applicants who use Chromebook computers.
Key skills/competency
- JSON
- Rubrics
- AI Model Grading
- Data Assessment
- Detail-Oriented
- Task Description Interpretation
- Consistency
- US-Based
- Part-Time
- Weekend Sprint
Skills & topics
- AI Model Grader
- JSON
- Rubrics
- Data Annotation
- Part-Time
- Weekend Work
- US-Based
- Remote
- Data Analysis
- AI Evaluation
How to get hired
- Tailor your resume: Highlight JSON experience and rubric application skills for the AI Model Grader role.
- Emphasize immediate availability: Clearly state your ability to start on July 24th and commit to the sprint.
- Showcase attention to detail: Provide examples of your consistent and accurate grading in past roles.
- Confirm technical setup: Mention you have a desktop or laptop computer, not a Chromebook.
Technical preparation
Practice parsing and understanding JSON structures.,Familiarize yourself with rubric-based assessment.,Review common AI model output formats.,Ensure reliable internet and computer access.
Behavioral questions
Describe a time you ensured grading consistency.,How do you handle ambiguity in instructions?,Detail your approach to meticulous data review.,How do you manage time for quick turnarounds?
Frequently asked questions
- What is the primary responsibility of an AI Model Grader at Mercor?
- As an AI Model Grader for Mercor, your primary responsibility is to assess AI model outputs against a defined rubric, ensuring consistency and accuracy in grading human data tasks. This involves understanding task descriptions, interpreting rubrics, and parsing JSON-formatted outputs.
- What is the expected time commitment for this AI Model Grader role?
- This is a part-time role requiring a minimum commitment of 20-40 hours per week. Tasks are expected to be completed within a 12-24 hour turnaround time after assignment.
- What technical requirements are needed for the AI Model Grader position?
- You will need access to a desktop or laptop computer. Chromebooks are not accepted for this role. Proficiency in reading and interpreting JSON files is also essential.
- What is the duration and start date for this AI Model Grader project?
- The project is a weekend sprint beginning immediately on July 24th and running through July 27-28. There is a possibility for extension.
- Does Mercor prefer candidates with prior experience for the AI Model Grader role?
- While prior experience with rubrics is a plus, Mercor particularly values candidates with prior Mercor experience, especially in applying detailed rubrics and evaluation criteria consistently. Experience processing JSON files is also highly beneficial.
- Are AI tools permitted while grading for Mercor?
- No, you must complete all assessments and grading tasks without the assistance of AI tools to ensure human judgment and consistency.
- What is the geographic requirement for this AI Model Grader position?
- Applicants for this role must be based in the US.