Software Engineer, Agentic Tooling, Tesla SI

Sourced from justjobs.com

Tesla SI's Agentic Tooling team builds the harness and skill ecosystem powering FSD and Optimus. You'll extend the agent runtime, ship new skills, build evaluation frameworks, and own observability for production agent services. The role is collaboration-heavy, embedding with engineers across the org to turn vague asks into shipped agent-powered tools.

• Extend the agent runtime and tool ecosystem end-to-end, adding tools, integrations, and backends
• Ship new skills that solve real engineering workflows across the SI org
• Build and operate the evaluation framework that catches regressions in prompts, skills, and model upgrades
• Own session-level observability across tracing, logging, and metrics; build dashboards and alerting
• Operate production agent services on Kubernetes with async backends and durable storage
• Contribute to architectural decisions around security, scalability, and reliability
• Embed with users and drive solutions from prototype through rollout and hand-off

• Proficiency with Python
• Strong foundation in concurrent, async, and distributed systems and orchestration
• Strong software-design fundamentals (SOLID, clean abstractions, composition)
• Hands-on experience building with agents (tools, skills, harnesses, or side projects)
• Ability to convert vague asks into shipped solutions end-to-end
• Adaptability and curiosity in a fast-moving space

• Go experience
• Familiarity with Docker, Kubernetes, relational and vector databases
• Experience with Claude Code SDK, ACP, or opencode
• Building or operating evaluation pipelines (offline replay, ground-truth scoring, CI gating)
• Experience with sandboxing and application security (container isolation, credential isolation, secrets management)