Computational Structural Engineering AI Benchmark Designer

About OpenTrain

OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. OpenTrain AI is hiring contractors to create specialized evaluation work that helps shape how advanced AI systems reason about real technical problems.

This opportunity is available worldwide as a part-time contractor engagement. Creating an OpenTrain account is free, and contributors can use the platform to build a profile and apply to AI training projects.

• Worldwide opportunity
• Contractor and part-time engagement
• Apply through OpenTrain

About AI Benchmark Design

AI training is the human side of building artificial intelligence. In this role, you will design and evaluate challenging computational tasks so AI models can be tested on authentic scientific software workflows rather than simple text exercises.

Your work will help measure whether models can execute complex multi-step calculations, plan efficient queries or experiments, and produce reliable technical results.

• Work at the intersection of engineering, software, and AI evaluation
• Create research-level tasks based on real computational workflows
• Contribute to the development of more capable AI systems

The Role

OpenTrain is seeking a Computational Structural and Mechanical Engineering Expert to design original benchmark problems for advanced AI models. Problems may require exact answers from fully defined setups or strategic planning to uncover hidden information through a sequence of queries or experiments.

Each benchmark is tested against state-of-the-art AI models. You will refine the problem design and difficulty based on test results and feedback, then implement the supporting technical components in Python.

• Hourly compensation: $70 to $85 USD
• Listed rate: $85 USD per hour
• At least 15 to 20 hours per week, with a default commitment of 40 hours per week
• English-language work
• Entry-level listing with graduate-level specialist requirements

What You'll Do

You will create computational problems that require skilled use of specialized open-source scientific software. The work involves building complete, testable evaluation tasks and adjusting them until they meet the intended challenge level.

You will also write the Python logic needed to establish expected results and assess model submissions, while working through testing loops with remote compute sandboxes.

• Create original graduate-level structural and mechanical engineering problems
• Design tasks involving complex, multi-step computational workflows
• Develop problems with exact answers from fully defined setups
• Design query or experiment planning tasks that use partial results strategically
• Write problem setups, oracle functions, and solution validators in Python
• Test benchmarks against advanced AI models
• Iterate on difficulty and problem design using evaluation feedback

Required Qualifications

You should have graduate-level training in a relevant STEM field, such as an MS, PhD, or equivalent research experience, with expertise in computational structural or mechanical engineering.

Strong programming ability and practical experience with scientific computing are essential. You should also be comfortable working independently in technical environments where problems and evaluation logic must be developed and refined.

• Graduate-level expertise in computational structural or mechanical engineering
• MS, PhD, or equivalent research experience in a relevant STEM field
• Hands-on proficiency with at least one listed open-source computational tool
• Strong Python skills for problem setups, oracle functions, and solution validators
• Ability to design challenging evaluation problems that test AI model reasoning
• Comfort working independently and incorporating feedback
• Experience with Linux, terminal tools, and remote compute sandboxes

Supported Scientific Tools

You must have proven hands-on proficiency with at least one open-source computational tool from the list below. Experience with multiple tools or domains is helpful but not required.

• FEniCSx or DOLFINx
• scikit-fem, deal.II, MFEM, or MOOSE
• CalculiX, Elmer FEM, Code_Aster, or SfePy
• FiPy, Devito, Cantera, CoolProp, Pyomo, or SimPy

Helpful Background

The strongest candidates may bring experience beyond the core requirements, particularly in benchmark development, scientific education, or reproducible computational research.

• Experience across multiple relevant computational domains or tools
• Familiarity with benchmark or evaluation design
• Scientific teaching or exam and problem-set design experience
• Experience with computational reproducibility
• Familiarity with containerized environments

How This Work Fits AI Careers

AI training and data labeling work includes evaluating model outputs, writing examples, and testing whether systems can solve demanding technical tasks. Unlike traditional annotation, this project calls for deep engineering knowledge and hands-on software development.

The work is remote-friendly and flexible, making it possible to contribute part time while applying specialized expertise to the fast-growing field of AI development.

• Use advanced engineering expertise to influence AI evaluation
• Work flexibly as a part-time contractor
• Build experience in a growing technical AI training field
• Apply through OpenTrain and develop your AI training career