Astrophysics Software Designer for AI Evaluation (Remote)

mercor; France; Remote; EUR 74,000 - 123,000; Full time

Job summary

Mercor is seeking a STEM Computational Scientific Software & Evaluation Designer in Astrophysics & Cosmology. This contract role focuses on crafting graduate-level computational problems using domain-specific libraries, and evaluating AI systems' ability to solve research-grade challenges.

You will work remotely 15–20 hours per week, using Python in Linux environments, with astropy and related tools for cosmological calculations and data reduction.

Qualifications

• Graduate-level STEM degree or equivalent research experience.
• Proficiency with scientific software libraries evidenced by publications or professional work.
• Strong Python programming skills for problem setups and validators.
• Ability to work in a Linux/terminal environment with remote compute sandboxes.
• Available for at least 15–20 hours per week.

Responsibilities

• Design graduate-level computational problems using domain-specific scientific software libraries.
• Evaluate AI systems' ability to solve research-grade problems, from querying simulations to interpreting outputs.
• Develop experimental strategies to recover hidden information from data, emphasizing strategic reasoning.
• Calibrate tasks against state-of-the-art AI models, refining problem designs to achieve target difficulty.
• Work independently and asynchronously to iterate on problem designs based on calibration feedback.
• Utilize astropy and related tools for cosmological calculations and observational data reduction.

Skills

Graduate STEM degree

Python programming

Scientific software libraries

Linux/terminal

Remote 15-20h/wk

Education

MS/PhD or equivalent

Tools

Astropy

Job description

Mercor is seeking a STEM Computational Scientific Software & Evaluation Designer in Astrophysics & Cosmology. This contract role focuses on crafting graduate-level computational problems using domain-specific libraries, and evaluating AI systems' ability to solve research-grade challenges.

You will work remotely 15–20 hours per week, using Python in Linux environments, with astropy and related tools for cosmological calculations and data reduction.