Biology AI Sci Code Architect for Benchmarks

Mercor; Toronto; On-site; CAD 34,000 - 62,000; Part time

Job summary

Mercor is seeking PhD or Master’s scientists to author executable AI evaluation tasks for a new scientific benchmarking effort. You will design original problems, source materials, and craft strict grading criteria to challenge frontier models.

The role emphasizes biology-focused domains with a strong emphasis on Python, R, and reproducible workflows using Git/GitHub and Docker. This six-week engagement invites part-time contributors to start immediately.

Qualifications

• PhD in biology or related field.
• Demonstrated depth in ecology, biochemistry, or genetics (at least two).
• Proficiency in Python, R, or another scientific computing language.
• Comfortable with Git/GitHub and running code in Docker via PR workflows.

Responsibilities

• Source your own material (papers, datasets, repositories, or designed scenarios).
• Write scientific prompts based on input material.
• Build grading criteria to define correct answers.
• Calibrate prompts against frontier models so tasks fail more often than succeed.

Skills

Python

R

Scientific computing

Git/GitHub

Docker

Education

PhD in biology or related field

Master's degree considered

Tools

Docker

Git/GitHub

Job description

Mercor is seeking PhD or Master’s scientists to author executable AI evaluation tasks for a new scientific benchmarking effort. You will design original problems, source materials, and craft strict grading criteria to challenge frontier models.

The role emphasizes biology-focused domains with a strong emphasis on Python, R, and reproducible workflows using Git/GitHub and Docker. This six-week engagement invites part-time contributors to start immediately.