Scientific Computing AI Evaluation Author

About OpenTrain

OpenTrain AI is the hiring and contracting organization for this role and the #1 platform for finding and building careers in AI training and data labeling. OpenTrain helps contributors discover specialized projects, build a credible profile, and grow their experience in a rapidly developing field.

Creating an OpenTrain account is free, and you can apply in minutes. This contract gives experienced scientific researchers an opportunity to contribute directly to how advanced AI systems are evaluated.

• Remote contract work with OpenTrain AI
• Build experience in scientific AI evaluation
• Create a lasting profile around specialized AI-training work

About AI Training and Scientific Evaluation

AI training is the human side of building modern artificial intelligence. Researchers and specialists prepare examples, write challenging prompts, and assess model responses so AI systems can produce more accurate and useful results.

In this role, your scientific expertise will help expose the limits of frontier models. You will turn research materials and original scenarios into executable problems with grading standards that clearly distinguish strong solutions from incorrect ones.

• Work at the intersection of physics, programming, and AI evaluation
• Help shape how advanced models solve scientific problems
• Contribute to cutting-edge AI development through structured human review

The Role

OpenTrain is seeking a Scientific Computing AI Evaluation Author to create original, executable research problems for evaluating frontier AI models. The work combines advanced physics knowledge, scientific programming, task design, and rigorous assessment of model performance.

You will work remotely on a six-week, part-time contract. The commitment is at least 20 hours per week, and the pay rate is $70 per hour.

• Contract length: Six weeks
• Schedule: Part time, at least 20 hours per week
• Location: Remote and worldwide
• Language: English
• Compensation: $70 per hour

What You'll Do

You will source reliable material and transform it into challenging scientific computing tasks. Each task should be executable, research-oriented, and supported by grading criteria that make evaluation consistent and meaningful.

You will also calibrate tasks against strong models and contribute through a GitHub pull-request workflow with automated quality checks.

• Source material from published papers, Kaggle datasets, open-source repositories, or original scenarios
• Write scientific prompts focused on executable research problems
• Define grading criteria that distinguish correct and incorrect answers
• Test and calibrate evaluation tasks against frontier models
• Use GitHub pull requests and automated quality checks
• Develop work across condensed matter, optics, quantum information and computing, computational physics, astrophysics, and particle physics

Requirements

This role requires PhD-level training and demonstrated research depth in physics, applied physics, or a closely related discipline. You should be able to formulate executable scientific computing problems and evaluate solutions using clear, technically rigorous standards.

You will also need practical experience with scientific programming and modern development workflows.

• PhD in physics, applied physics, or a closely related field
• Research depth in at least two relevant physics subdomains
• Ability to formulate executable scientific computing problems
• Ability to create rigorous grading criteria
• Working proficiency in Python, R, or another scientific programming language
• Practical GitHub pull-request experience
• Experience running code in Docker

Helpful Background

The strongest applicants will bring evidence of research communication and technical implementation in addition to formal physics training. These experiences can help you design realistic scientific tasks and assess whether a model's answer is genuinely correct.

• Peer-reviewed publications
• Scientific software development experience
• Research engineering experience

How to Apply Through OpenTrain

Create or update your free OpenTrain profile and apply for this contract in minutes. Highlight your physics research depth, scientific programming experience, relevant publications, and familiarity with GitHub and Docker.

OpenTrain brings together opportunities in AI training and data labeling so specialists can build a durable portfolio while contributing to the development of state-of-the-art AI systems.

• Create a free OpenTrain account
• Showcase your physics and scientific computing background
• Apply remotely for the six-week contract
• Be prepared to commit at least 20 hours per week