Applied Research, Chip Design

This is a fully remote position, open to applicants in United States, +2 more locations.

📋 Description

• Develop agents, environments, evaluations, and post-training systems tailored for semiconductor design and verification workflows.

• Transform workflows into distinct tasks, environments, evaluations, and reward signals.

• Conduct post-training experiments utilizing actual baselines.

• Ascertain whether improvements stem from the model, data, harness, inference, or infrastructure.

• Create interactive verification environments where agents can read specifications, modify or repair RTL, engage simulation or formal tools, interpret failures, and iterate towards verified outcomes.

• Design rewards and evaluations that integrate functional correctness, tool efficiency, robustness, and performance/power/area constraints.

• Train long-horizon agents to recover from design and tool failures, maintain useful states, and generalize across various blocks, designs, and constraint sets.

• Develop interactive environments enabling agents to utilize EDA tools for design modifications, debugging failures, and planning multi-step tasks.

• Create verifiers and rewards based on tests, simulations, formal methods, design rules, and engineering results.

• Experiment with credit assignment in scenarios where feedback is delayed and multiple tools are involved.

• Establish evaluations focused on correctness, recovery, and generalization.

• Collaborate directly with customers, users, or research partners while sustaining a meaningful research agenda.

⛳️ Requirements

• Strong background in reinforcement learning (RL), agents, or post-training systems, with the capacity to become proficient in semiconductor workflows.

• Previous experience in design automation, compilers, formal methods, or machine learning for systems is beneficial but not mandatory.

• Practical experience in constructing and troubleshooting model-training, post-training, agent, evaluation, or closely related experimental processes.

• Demonstrated experimental judgment: establishing meaningful baselines, meticulous trace analysis, clear failure attribution, and readiness to change course when evidence warrants.

• Proficient in Python.

• Capability to create artifacts that can be executed, inspected, and extended by other researchers and engineers.

• Comfort in addressing ambiguous challenges and keeping partners and teammates informed.

• Proven record of outstanding work, demonstrated through publications, open-source contributions, benchmarks, or delivered products.

🏝️ Benefits

• Equity incentives.

• Flexible work options (San Francisco, New York, London, or remote).

• Visa sponsorship and relocation assistance.

• Professional development budget.

• Team off-sites and conference participation.