Frontier AI Applied Researcher - Regulated Systems (Remote)

Prime Intellect, Inc.; United States; Remote; USD 150.000 - 300.000; Vollzeit

Benefits dieser Stelle

Equity incentives

Remote work option

Visa sponsorship & relocation

Professional development budget

Team off-sites & conferences

Zusammenfassung

Prime Intellect, Inc. is building open frontier AI and a unified platform for post-training at frontier scale. You will work on agent systems in public-sector and regulated industries, collaborating with customers to deploy robust AI solutions.

This applied research role emphasizes reusable research, evaluations, and safety, with opportunities to publish and contribute to open-source components. Expect cross-functional collaboration within boundary constraints.

Qualifikationen

• Experience with RL/post-training, private ML systems, and high-assurance evaluation.
• Hands-on ability to build and debug model-training, post-training, or related loops.
• Strong Python and ability to produce runnable artifacts for others.

Aufgaben

• Turn workflows into tasks, environments, evaluations, and reward signals and deploy effective results.
• Run post-training experiments against baselines; determine if gains come from model, data, or infrastructure.
• Develop reusable research components and contribute to open-source systems and products.
• Collaborate with customers, users, or partners; work within partner constraints and access rules.
• Create high-assurance evaluations for reliability, calibration, robustness, and compliance.

Kenntnisse

Python

Reinforcement Learning

Agent safety

Experimental judgment

Public-sector tech

Jobbeschreibung

Prime Intellect, Inc. is building open frontier AI and a unified platform for post-training at frontier scale. You will work on agent systems in public-sector and regulated industries, collaborating with customers to deploy robust AI solutions.

This applied research role emphasizes reusable research, evaluations, and safety, with opportunities to publish and contribute to open-source components. Expect cross-functional collaboration within boundary constraints.