Applied RL for Chip Design & Verification

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

Benefits offered by this job

Professional Development budget

Equity incentives

Remote work options

Job summary

Prime Intellect, Inc. is developing an open frontier AI platform and applying RL-driven agents to semiconductor design and verification workflows.

You will create interactive environments, evaluations, and post-training systems that operate with real toolchains and domain experts, while maintaining a practical research agenda and external collaborations. The role is externally facing, with opportunities to work with customers and partners, and to contribute reusable methods and open-source

Qualifications

• Depth in RL, agents, or post-training systems and fluency in semiconductor workflows.
• Hands-on experience building and debugging model-training, post-training, agent, evaluation loops.
• Strong Python and ability to build artifacts that researchers/engineers can run and extend.
• Excellent experimental judgment and willingness to pivot when evidence changes.

Responsibilities

• Build agents, environments, evaluations, and post-training systems for semiconductor design workflows.
• External-facing role with customers, users, or partners.
• Design rewards and evaluations balancing correctness with tool efficiency and resource constraints.
• Train long-horizon agents to recover from design/tool failures and generalize across blocks.

Skills

Reinforcement Learning

Agent systems

Post-training systems

Python

Tools

EDA tools

Job description

Prime Intellect, Inc. is developing an open frontier AI platform and applying RL-driven agents to semiconductor design and verification workflows.

You will create interactive environments, evaluations, and post-training systems that operate with real toolchains and domain experts, while maintaining a practical research agenda and external collaborations. The role is externally facing, with opportunities to work with customers and partners, and to contribute reusable methods and open-source