Astrophysics AI Benchmark Architect

Obsidian; San Francisco (CA); On-site; USD 120,000 - 180,000; Full time

Job summary

Obsidian is seeking a Computational Astrophysics & Cosmology Expert to design graduate-level problems that test AI systems using real scientific software, including simulations and data analysis in cosmology workflows.

You will develop challenging problems, test them against advanced AI models, and refine them to hit target difficulty. Strong Python skills, hands-on Astropy experience, and Linux familiarity are essential; 15–20 hours per week work is expected, with remote compute sandbox access.

Qualifications

• Graduate-level training in a relevant STEM field (MS/PhD preferred).
• Proven proficiency with at least one scientific software library (e.g., Astropy).
• Strong Python skills for writing problem setups, oracle functions, and validators.
• Ability to work independently and refine problem designs based on feedback.
• Comfortable working in Linux/terminal environments with remote compute sandboxes.
• Available for 15–20 hours per week.

Responsibilities

• Design graduate-level computational problems using real scientific workflows.
• Test problems against state-of-the-art AI models and refine difficulty.
• Create problems that require strategic planning and multi-step workflows.

Skills

Python

Astrophysics knowledge

Linux/Unix

Education

MS/PhD in a relevant STEM field

Tools

Astropy

Job description

Obsidian is seeking a Computational Astrophysics & Cosmology Expert to design graduate-level problems that test AI systems using real scientific software, including simulations and data analysis in cosmology workflows.

You will develop challenging problems, test them against advanced AI models, and refine them to hit target difficulty. Strong Python skills, hands-on Astropy experience, and Linux familiarity are essential; 15–20 hours per week work is expected, with remote compute sandbox access.