About The Position

Patronus AI is hiring an independent contractor to create reinforcement learning tasks related to embedded firmware that test and train AI models. The work is remote and performed on the contractor’s own schedule, with a commitment of 10 to 30 hours per week. The role involves designing realistic tasks that take an experienced engineer about one working day, building the sandboxed setup needed for each task, developing a solution and objective checks that a tester can verify, and reviewing tasks created by other experts. Setups may involve C and C++, RTOS, device drivers, bootloaders, and hardware-in-the-loop tests. All systems and scenarios must be fictional. Work must not be copied from published sources or use code, configurations, or data from a current or former employer or client. AI tools may be used as an aid, but contractors are responsible for their own work and are evaluated on it. Compensation is $150 per hour, and the start is rolling. Hiring includes a short AI interview and resume review, one paid trial task scored against a fixed quality checklist, and onboarding followed by a first project.

Requirements

  • At least 3 years writing production firmware for microcontrollers or embedded Linux devices.
  • Previous AI data labeling, data annotation, or reinforcement learning task creation experience.
  • Must be currently in the role or have left recently.
  • Responsibility for and evaluation of one's own work, even when AI tools are used as an aid.

Nice To Haves

  • Managerial experience.

Responsibilities

  • Design realistic reinforcement learning tasks related to embedded firmware to test and train AI models.
  • Build the sandboxed setup needed for each task, which may include C and C++, RTOS, device drivers, bootloaders, and hardware-in-the-loop tests.
  • Develop solutions and checks that testers can verify objectively.
  • Review tasks created by other experts.
  • Create fictional systems and scenarios without copying published sources or using employer or client code, configurations, or data.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service