Job description

About Patronus AI Patronus AI is a frontier lab developing simulation research and infrastructure to accelerate progress toward human-aligned AGI. We are on a mission to simulate all of the world’s intelligence. We are the team behind some of the earliest and most influential research in AI evaluation like FinanceBench , Lynx , SimpleSafetyTests , CopyrightCatcher , Humanity’s Last Exam , and more. We are formerly AI researchers and engineers from companies like Meta AI, Amazon AGI, and Google. Our customers include foundation model labs and Fortune 500 enterprises like Adobe. We are backed by top-tier investors like Lightspeed Venture Partners, Notable Capital, Stanford University, Noam Brown, Gokul Rajaram, and more. At a glance Engagement: Independent contractor, fully remote, asynchronous Pay: $50 to $150 per hour. Some projects pay per task instead: you are paid when the full task is submitted, with a bonus for each accepted task. Strong performers move up within that range or onto productivity-based pay. The trial task is paid. Commitment: 10 or more hours per week Start: Rolling Location: Remote The role You'll work with our research team on two things. The first is helping us understand how work in your area is actually done. The second is designing realistic mechanical engineering tasks that a qualified professional could solve with the right knowledge and tools, and that today's AI models get wrong. Each task starts from a scenario we provide. You create the working documents someone in your role would handle, write the request, solve it yourself and define how the answer is graded. Your work becomes evaluation and training data for frontier models. You don't need an AI background. You do need to be the kind of mechanical engineer who checks the load case before trusting the safety factor. Areas we're hiring for Each area is defined by what governs the work, so you should fit exactly one. If you work across two, apply under the one where you spend most of your time. Mechanical design and structural analysis. Design and stress engineers who turn requirements into parts, assemblies and mechanisms governed by stress, deflection, fatigue and motion. They own machine elements, GD&T and tolerancing, hand calculations and FEA, drawings and bills of materials, and the structural and vibration tests that prove them. If heat or flow governs the sizing, the work belongs to thermal and fluid systems. Typical titles: Mechanical Design Engineer, Product Design Engineer, Mechanical Engineer, Stress Analyst, FEA Engineer, Structural Analyst. Thermal and fluid systems. Thermal and fluids engineers who size and specify systems governed by heat transfer and flow: HVAC and building mechanical systems, thermal management for electronics and batteries, piping, pumps and heat exchangers, and turbomachinery. They own load and flow calculations, equipment selection, CFD and the thermal and flow tests that prove them. The structure that carries the system belongs to design. Typical titles: HVAC Engineer, Building Mechanical Engineer, Thermal Engineer, Fluids Engineer, CFD Engineer, Piping Engineer. Manufacturing processes and tooling. Manufacturing engineers who decide how mechanical parts are made: process selection, design for manufacturing and assembly, tooling, fixtures, molds and dies, materials selection and heat treatment, additive manufacturing and process capability. Whether a part can be made belongs here. Whether it should be designed differently belongs to design. Chemical process engineering and production line optimization are not included. Typical titles: Manufacturing Engineer, Tooling Engineer, Tool and Die Designer, Mold Designer, Additive Manufacturing Engineer, Process Development Engineer. What you'll do Build realistic evidence files from a provided scenario, such as drawings and CAD models, bills of materials, hand calculations and simulation reports, load calculations and equipment schedules, process sheets and tooling drawings, and test reports. Write task requests that state the deliverable and say how uncertainty and justified assumptions should be handled. Solve each task end to end and document a reference solution with the reasoning behind every step. Write scoring criteria precise enough that two independent graders reach the same verdict, including acceptable alternative answers and tolerances on numerical results. Check your own and your peers' tasks for technical accuracy and ambiguity. Work in reviewer feedback within agreed turnaround times, and flag guideline gaps or edge cases you run into. Explain to our research team how work in your area is actually done: the steps, the documents, who does what, and where things go wrong. Review the scenarios and documents our team builds from that research, and flag anything a practitioner would find unrealistic. Who we're looking for Required 3+ years of hands-on experience in at least one of the areas above, as a mechanical design or stress engineer, a thermal, fluids or HVAC engineer, or a manufacturing or tooling engineer. You did the work yourself; supervising it is not enough. Working command of the standards, methods and codes in your area. Examples: ASME Y14.5 and GD&T, hand-calculation and finite element methods, ASHRAE and ASME codes, heat transfer and fluid sizing methods, or design for manufacturing and process capability. Strong calculation and spreadsheet skills and the habit of checking every unit, assumption and load case. Clear, precise writing in professional English. Ability to work independently and asynchronously against deadlines. Nice to have Prior experience writing or reviewing AI evaluation tasks, exam questions, case studies or training material. Familiarity with how AI models fail. Our standards All tasks, documents and solutions are human-authored. LLM-generated content is not permitted. Tasks and solutions must be original. Nothing may be copied from published or easily searchable sources. Never use real employer, client or supplier data, drawings or designs, or anything covered by an NDA or export control. All products, companies and scenarios are fictional. Quality is judged on accuracy, rubric validity, difficulty and reviewer agreement, not volume. How hiring works Short AI interview and resume review One paid trial task, scored against a fixed quality checklist Onboarding and your first project Patronus AI is an equal opportunity employer. We celebrate diversity in our workplace, and all qualified applicants will receive consideration for employment without regard to age, ancestry, color, family or medical care leave, gender identity or expression, genetic information, marital status, medical condition, national origin, physical or mental disability, political affiliation, protected veteran status, race, religion, sex (including pregnancy), sexual orientation, or other legally protected characteristics. By clicking ‘Apply’, you agree to both of the following: 1) Greenhouse's Terms of Service and Privacy Policy 2) Patronus AI, Inc. Privacy Policy .

Requirements

Must-haves

You don't need an AI background. You do need to be the kind of mechanical engineer who checks the load case before trusting the safety factor.
Areas we're hiring for
Each area is defined by what governs the work, so you should fit exactly one. If you work across two, apply under the one where you spend most of your time.
Write task requests that state the deliverable and say how uncertainty and justified assumptions should be handled.
Solve each task end to end and document a reference solution with the reasoning behind every step.
Write scoring criteria precise enough that two independent graders reach the same verdict, including acceptable alternative answers and tolerances on numerical results.
Check your own and your peers' tasks for technical accuracy and ambiguity.
Work in reviewer feedback within agreed turnaround times, and flag guideline gaps or edge cases you run into.
Explain to our research team how work in your area is actually done: the steps, the documents, who does what, and where things go wrong.
Review the scenarios and documents our team builds from that research, and flag anything a practitioner would find unrealistic.
Who we're looking for
Strong calculation and spreadsheet skills and the habit of checking every unit, assumption and load case.

Nice-to-haves

Prior experience writing or reviewing AI evaluation tasks, exam questions, case studies or training material.
Familiarity with how AI models fail.
Our standards
All tasks, documents and solutions are human-authored. LLM-generated content is not permitted.
Tasks and solutions must be original. Nothing may be copied from published or easily searchable sources.
Never use real employer, client or supplier data, drawings or designs, or anything covered by an NDA or export control. All products, companies and scenarios are fictional.
Quality is judged on accuracy, rubric validity, difficulty and reviewer agreement, not volume.
How hiring works
Short AI interview and resume review
One paid trial task, scored against a fixed quality checklist
Onboarding and your first project
By clicking ‘Apply’, you agree to both of the following:

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

aigollm

Platform: greenhouse
Effort: quick