Scientific Coding: Physics and Python

What you'll do

You build benchmark problems out of physics. Each task is one main problem with at least three sub-problems that build logically toward it, plus a verified "golden" Python solution with full unit test coverage. The tests have to be discriminative, meaning they separate correct model output from plausible but wrong output.

Tasks go through quality control on Turing's Central Task Platform: Tier 1 structure checks and Tier 2 quality rubrics. You iterate on feedback until tasks meet Pass@K criteria across several LLM judges (the ad names GPT, Gemini and Nemotron), aiming for approval on first submission with little rework. You also join sync calls for reviews, feedback and standups during overlap hours.

Who qualifies

• Master's or PhD in Physics
• Strong Python with scientific computing experience
• Skill at writing rigorous, well-posed problems with clear constraints and expected outputs
• Attention to detail: tasks are held to rubrics for well-posedness, test discriminativeness, scientific correctness and determinism
• Prior AI data annotation, research or scientific writing experience
• Familiarity with LLM evaluation frameworks or coding benchmarks, and libraries like NumPy, SciPy or SymPy
• Published research or academic project experience in a STEM domain

What it pays

The ad does not publish a rate. Turing sets rates per project, see Turing.

Worth knowing

Good:

• Physics is the only discipline here that asks specifically for a physics degree, so it suits those who want to use it directly
• Clear quality bar, so you know what "done" looks like

Less good:

• 40 hours a week with 4 hours of Pacific-time overlap for 8 weeks
• Contractor assignment with no medical or paid leave
• Location limited to Bangladesh, Brazil, Colombia, Egypt, Ghana, India, Pakistan, Indonesia, Kenya, Nigeria, Turkey and Vietnam
• Rework against QC rubrics can eat time

About this listing

Posted on Turing's job board and read on 18 October 2026; the ad is one of several SciCode roles split by discipline. Terms are Turing's; commentary is ours. See Turing.