Scientific Computing Expert — Physical Sciences

Turing, a leading company in AGI infrastructure, collaborates with cutting-edge AI labs to accelerate model development by providing high-quality training data, evaluations, and engineering expertise. We are currently looking for experts in Physical Sciences to contribute to Terminal Bench Science by developing realistic, terminal-based scientific tasks. This role involves translating authentic workflows from physics, chemistry, materials science, astronomy, and computational science into reproducible benchmark environments. The goal is to help evaluate AI agents’ ability to reason through scientific problems, operate command-line tools, debug calculations, and generate reliable scientific outputs.

Key Responsibilities
Design and develop multi-step terminal tasks that reflect realistic physical-science workflows. Create self-contained computational environments with pinned dependencies and relevant scientific software. Generate input datasets such as molecular structures, simulation parameters, experimental data, or model configurations. Implement expert solutions using programming languages like Python, Bash, C/C++, Julia, or domain-specific tools, covering areas such as simulations, numerical modeling, data fitting, optimization, spectroscopy, molecular analysis, and scientific visualization. Develop automated tests to assess numerical accuracy, physical consistency, convergence, and output structure, defining appropriate numerical tolerances, units, boundary conditions, and expected scientific behavior. Ensure tasks are reproducible and execute without requiring runtime downloads. Troubleshoot and debug issues related to dependencies, numerical precision, solver stability, performance, and file formats.

Required Qualifications
A Ph.D., postdoctoral experience, or equivalent advanced technical expertise in a relevant physical-science discipline is essential. Strong programming skills in Python, C/C++, Julia, Bash, or other scientific programming languages are required. Experience working in Linux or terminal-based environments is necessary. Candidates should have a solid background in numerical methods, scientific modeling, simulations, or quantitative data analysis. The ability to independently create and validate computational scientific workflows is critical.

Engagement Details
This is a contractor assignment requiring a minimum commitment of 6 hours per day and at least 40 hours per week, with a mandatory 4-hour overlap with Pacific Standard Time (PST). The contract duration is 5 weeks, with an expected start date in the coming week. Compensation is task-based, with a payment of $300 per approved task. Please note that this is a contractor role without medical benefits or paid leave.

This opportunity is ideal for physical-science professionals who enjoy applying their expertise to develop rigorous scientific benchmarks and contribute to advancing AI capabilities in scientific reasoning and computation.