Remote Frontier AI Applied Researcher

Prime Intellect AI; San Francisco (CA); Remote; USD 150,000 - 300,000; Full time

Benefits offered by this job

Cash compensation: $150k–$300k + stock

Flexible work locations (SF, NYC, LDN,

Remote work option

Visa sponsorship & relocation

Professional development budget

Team off-sites & conference attendance

Job summary

Prime Intellect AI is building a post-training, agent-based platform for frontier AI within strict data and governance boundaries. You will design and implement end-to-end improvement loops, tool-using agents, and rigorous evaluations to ensure reliability, calibration, and compliance in regulated environments.

We value open research, reproducibility, and artifacts that other researchers can run and extend. Join a team pushing the boundaries of responsible frontier AI at scale.

Qualifications

• Experience in RL/post-training and private ML systems.
• Hands-on in model-training, post-training, agent, or evaluation loops.
• Ability to build artifacts for researchers to run and inspect.
• Strong experimental judgment and ethical compliance.
• Comfort with owning ambiguous problems and partner communication.

Responsibilities

• Build end-to-end improvement loops inside operator boundaries.
• Develop agents with permissions, evidence trails, bounded tools.
• Create evaluations for reliability, calibration, and compliance.
• Run experiments on privacy and infrastructure constraints.

Skills

RL/post-training

Python

Experimentation

Public-sector ML

Agent safety

Job description

Prime Intellect AI is building a post-training, agent-based platform for frontier AI within strict data and governance boundaries. You will design and implement end-to-end improvement loops, tool-using agents, and rigorous evaluations to ensure reliability, calibration, and compliance in regulated environments.

We value open research, reproducibility, and artifacts that other researchers can run and extend. Join a team pushing the boundaries of responsible frontier AI at scale.