Safety Engineer Intern: Lean Verified Autonomy Proofs

We Love AI Jobs; San Francisco (CA); On-site; USD 28,000 - 41,000; Part time

Job summary

Bedrock Robotics in San Francisco, CA invites an Engineering intern to apply safety claims for autonomous excavators using formal methods and Lean verification. You will formalize a claim, verify a Lean checker, and run it on fleet data, documenting assumptions and auditable results.

You'll collaborate with Safety, Systems, and Autonomy teams to improve safety case processes, record proof dependencies, and develop end-to-end demonstrations for one claim, contributing to auditable evidence and

Qualifications

• Hands-on experience with Lean and mathlib in real analysis or probability.
• Working knowledge of probability and statistics.
• Familiarity with knowledge graphs or structured document analysis.
• Ability to translate informal claims into precise specifications and identify their assumptions.
• Experience evaluating LLM-generated code or proofs.
• Background in autonomous vehicles, heavy equipment, or other safety-critical robotics.

Responsibilities

• Formalize a safety claim, implement and verify a Lean checker, and evaluate it on fleet data.
• Develop auditable checks for assumptions and document evidence that requires other sources.
• Review specifications with Safety, Systems, and Autonomy teams and recommend safety case improvements.

Skills

Lean

mathlib

probability

statistics

LLM proofs

knowledge graphs

fleet data

Rust verification tools

Tools

Aeneas

Verus

Kani

Job description

Bedrock Robotics in San Francisco, CA invites an Engineering intern to apply safety claims for autonomous excavators using formal methods and Lean verification. You will formalize a claim, verify a Lean checker, and run it on fleet data, documenting assumptions and auditable results.

You'll collaborate with Safety, Systems, and Autonomy teams to improve safety case processes, record proof dependencies, and develop end-to-end demonstrations for one claim, contributing to auditable evidence and