Autonomous Scientific AI Researcher

Prime Intellect, Inc.; New York (NY); Remote; USD 150,000 - 300,000; Full time

Benefits offered by this job

Flexible work locations

Professional development budget

Equity incentives

Job summary

Prime Intellect, Inc. is building the open superintelligence stack for frontier AI labs. Our Lab platform unifies compute, environments, evaluations, secure sandboxes, and deployment for post-training at frontier scale.

We are developing open frontier AI with end-to-end training of open-source models for long-horizon tasks. You will build and post-train agents for scientific research, spanning literature search, code execution, domain tools, and experiment selection, with environments tied to

Qualifications

• Depth in RL, agents, or post-training systems and ability to acquire scientific context.
• Experience in autonomous research, ML for science, or reproducible computational research is useful but not required.
• Hands-on experience building and debugging model-training, post-training, agent, evaluation loops.

Responsibilities

• Build and post-train agents for scientific research across literature, code and domain tools.
• Design long-horizon environments and experimental loops with research partners' workflows.
• Turn partner datasets and criteria into trainable tasks, evaluations, and reward signals.
• Develop RL and multi-agent methods and compare with baselines under fixed compute.
• Keep code, data, logs, and configurations reproducible and release reusable environments.

Skills

Reinforcement Learning

Agent-based systems

Post-training evaluation

Python

Tools

Python tooling

Job description

Prime Intellect, Inc. is building the open superintelligence stack for frontier AI labs. Our Lab platform unifies compute, environments, evaluations, secure sandboxes, and deployment for post-training at frontier scale.

We are developing open frontier AI with end-to-end training of open-source models for long-horizon tasks. You will build and post-train agents for scientific research, spanning literature search, code execution, domain tools, and experiment selection, with environments tied to