
51 - 200 employees
Founded 2023
🤖 Artificial Intelligence
🔬 Science
🔥 Funding within the last year
💰 $50M Series B on 2026-06
Artificial Intelligence • Science
Patronus AI is a behavioral research lab for artificial intelligence that studies where models and agents fail and how to improve their judgment. They build realistic, domain-specific simulated environments and world models where agents can practice decision-making under fragmented information, conflicting requests, and unstated requirements; analyze agent mistakes; and create training data, benchmarks, and tests (e. g. , TRAIL, MEMTRACK) to improve AI reliability and alignment. The company emphasizes fundamental research, publishes on arXiv, and provides resources such as RL environments, testing frameworks, and domain guides. Their site references enterprise-relevant compliance (SOC 2, HIPAA, TISAX).
🔥 46 minutes ago
🇺🇸 United States – Remote
💵 $50 - $150 / hour
⏳ Contract/Temporary
🟡 Mid-level
🟠 Senior
🧑🔧 Mechanical Engineer
👻 Ghost score 0%
Where this job comes from
Posted on Patronus AI's careers page, not through a job board
First seen by Remote Rocketship 46 minutes ago

51 - 200 employees
Founded 2023
🤖 Artificial Intelligence
🔬 Science
🔥 Funding within the last year
💰 $50M Series B on 2026-06
Artificial Intelligence • Science
Patronus AI is a behavioral research lab for artificial intelligence that studies where models and agents fail and how to improve their judgment. They build realistic, domain-specific simulated environments and world models where agents can practice decision-making under fragmented information, conflicting requests, and unstated requirements; analyze agent mistakes; and create training data, benchmarks, and tests (e. g. , TRAIL, MEMTRACK) to improve AI reliability and alignment. The company emphasizes fundamental research, publishes on arXiv, and provides resources such as RL environments, testing frameworks, and domain guides. Their site references enterprise-relevant compliance (SOC 2, HIPAA, TISAX).
• Help the research team understand how mechanical engineering work is performed in practice • Design realistic mechanical engineering tasks that qualified professionals can solve and current AI models get wrong • Create working documents such as drawings, CAD models, bills of materials, hand calculations, simulation reports, load calculations, equipment schedules, process sheets, tooling drawings, and test reports • Write task requests specifying deliverables and how uncertainty and justified assumptions should be handled • Solve tasks end to end and document reference solutions with reasoning • Write precise scoring criteria, including acceptable alternatives and numerical tolerances • Check personal and peer tasks for technical accuracy and ambiguity • Incorporate reviewer feedback within agreed turnaround times and flag guideline gaps or edge cases • Explain workflows, documents, responsibilities, and failure points to the research team • Review research-created scenarios and documents and flag anything unrealistic to practitioners
• 3+ years of hands-on experience in mechanical design or structural analysis, thermal/fluid/HVAC engineering, or manufacturing/tooling engineering • Working command of relevant standards, methods, and codes, including examples such as ASME Y14.5 and GD&T, ASHRAE and ASME codes, heat transfer and fluid sizing methods, or design for manufacturing and process capability • Strong calculation and spreadsheet skills • Habit of checking every unit, assumption, and load case • Clear, precise writing in professional English • Ability to work independently and asynchronously against deadlines • Human-authored work only; LLM-generated content is not permitted • Original tasks and solutions; no copying from published or easily searchable sources • No use of real employer, client, or supplier data, drawings, or designs, or anything covered by an NDA or export control • Applicants must complete a short AI interview, resume review, and one paid trial task • Prior AI evaluation, exam question, case study, or training-material experience is nice to have • Familiarity with how AI models fail is nice to have
• Pay of $50 to $150 per hour • Bonus for each accepted task for projects paid per task • Strong performers may move up within the pay range or onto productivity-based pay • Paid trial task • Fully remote, asynchronous work • 10 or more hours per week • Rolling start
Apply NowImprove your chances of getting an interview by checking your resume score before you apply.
🕒 6 days ago
Mechanical Engineer developing and integrating balance-of-site systems for NuScale’s small modular reactor power plants. Providing technical oversight to design partners and resolving complex engineering challenges.
🕒 October 20
51 - 200
Mechanical and CAD engineers completing narrated design workflows for Mercor. Creating high-quality training data to benchmark and improve frontier AI models.
🇺🇸 United States – Remote
💵 $50 - $60 / hour
🔥 Funding within the last year
💰 $350M Series C - Mercor on 2025-10
⏳ Contract/Temporary
🟢 Junior
🟡 Mid-level
🧑🔧 Mechanical Engineer
🕒 October 20
Mechanical Engineer leading ASME Section III stress analysis of pressure vessels and supports. Supporting NuScale’s modular advanced light water reactor systems through technical leadership and engineering oversight.
🕒 October 19
Mechanical Data Engineer creating CAD-grounded datasets, annotations, and quality workflows for an AI copilot startup. Applying manufacturing expertise to improve model training, evaluation, and customer delivery.
🕒 October 13
51 - 200
CAD, PCB, and 3D engineers creating realistic design tasks and automated graders for Mercor’s frontier AI evaluation projects. Testing models and refining engineering benchmarks.
🇺🇸 United States – Remote
💵 $60 - $100 / hour
🔥 Funding within the last year
💰 $350M Series C - Mercor on 2025-10
⏳ Contract/Temporary
🟡 Mid-level
🟠 Senior
🧑🔧 Mechanical Engineer