Help train a leading foundational AI lab’s language model on professional presentation work. You’ll create, assess, and refine AI-generated slide decks for technical sales workflows, drawing on your experience building presentations as a sales engineer or solutions consultant. The role is fully remote, offers schedule flexibility, and has variable client-driven hours; most experts work fewer than 10 hours per week, with no weekly cap.
Turn engineering computations you have implemented and tested into research-grade tasks for AI model evaluation. Each task centers on a real calculation from your research, with a main problem, related subproblems, a Python reference solution, and tests. You can work around your own schedule, submit tasks for review, and choose whether to assess submissions from other experts.
Help assess and improve coding agents by giving them real engineering tickets, running their work, and tracing mistakes through a codebase you know. The contract offers 40 hours each week at $6,000, paid each week, with hours you set. It renews each week when your work meets quality standards. Selection involves a proctored coding assessment and a paid trial task.
Build and refine software engineering evaluations from completed pull requests, using Mercor’s in-house framework to assess frontier coding agents such as Claude Code and Codex. You’ll work with an internal research team whose benchmarks compare frontier large language models, then take ownership of evaluation selection, development, and refinement after onboarding.
AfterQuery engages licensed engineers to create and assess scenarios that test how AI models handle professional engineering decisions. You will build examples spanning design, review, and approval, then explain the code and safety considerations behind sound answers. The work is remote and asynchronous, with a minimum commitment of 10 hours per week.
Use your drilling and energy operations experience to create expert training data and assess AI-generated answers for accuracy and relevance. The work draws on knowledge of drilling workflows, well planning, and energy data tools. It is remote and project-based, with a flexible schedule and no minimum hours. Contributors often work 5–20 hours per week during active projects; future placements depend on project needs. No AI experience is required.
Apply to Mercor’s network for potential contract projects with AI labs and companies. After you complete your profile and pass an AI interview, Mercor may match you with projects as they open. Work can include training and evaluating AI models, developing frontend tasks based on real scenarios, and giving domain feedback for AI research. Projects vary in length and scope, with commitments often around 15–30 hours per week.
AfterQuery needs construction professionals with 1 to 3 years of experience for project-based work over the next 2 to 3 weeks. You’ll review construction documents, support planning and field problem-solving, and evaluate technical questions across different project types. Work runs asynchronously, with a 10 to 20 hour weekly commitment for each project you accept.
Review coding agents as they tackle real engineering tickets, then work through problems in codebases you know to pinpoint and correct model mistakes. The contract offers 40 hours of work each week, with flexible scheduling and weekly renewal when your work meets quality standards. The selection process includes a proctored coding assessment and a paid trial task.
A senior engineering specialist will help a leading AI lab assess and improve how its models handle real software and engineering work. You will review model outputs, write task instructions and reference solutions, and create evaluation benchmarks with the research team. The position is full-time W-2 employment through Cincinnatus LLC, based in the Bay Area with on-site work multiple days each week when required.
Build substantial software projects that mirror real engineering work. You’ll shape multi-file codebases, create development challenges with clear documentation, and write deterministic tests that verify correct, repeatable results. The role calls for full-stack experience, strong technical writing, and hands-on use of Docker or a similar container tool.
Use your software engineering experience to assess coding agents on real engineering problems. You’ll give an agent a ticket and review its work, then work through a problem in a codebase you know and mark where the model fails. The contract offers 40 hours of work each week, with flexible scheduling and weekly renewal based on quality standards.
Review coding sessions created with AI developer tools for a frontier AI lab’s model training and evaluation work. You’ll judge whether each trace reaches a correct result, follows a sound development process, and shows sound reasoning. Your written feedback should explain your assessment and follow the provided rubric.
Join a research cohort of senior experts who assess how frontier AI models handle complex, real-world problems in hardware design, including CAD, RTL, and robotics. You will define standards for correct, high-quality work through scenario design, reference solutions, rubrics, and model-output reviews. The role centers on research and evaluation, not production engineering. The process includes a thorough background check.
Review coding agents on real engineering tickets, run their code, and assess the results as you would a teammate’s pull request. You’ll also work through a problem in a codebase you know, noting where the model falters. The contract provides 40 hours of work each week, with flexible scheduling, and requires at least four years of professional software engineering experience.
Location and eligibility
Remote · Eligibility not specified
Engagement type
Part-time
Compensation
$120/hr
software engineering
code review
coding agents
Software EngineeringView job
Showing 20 active roles of 58. Dates reflect what the source published, not when AIWorkJobs last synced.
Sources: ten specialist AI work platforms. AIWorkJobs does not verify every claim made by a source. Report a listing that looks wrong through the contact page.