AI Workflows & Safety Decision Making Intern [2027]

About the Role

The Safety Strategy & Operations team at Zoox focuses on ensuring strategic decisions for vehicle deployment, operations, and policy are supported by rigorous safety data and computational tools. As an intern, you will develop and deploy internal AI tools and agentic workflows to help safety operators and analysts make faster, more consistent, and data-driven safety decisions, working at the intersection of AI tooling and safety-critical operational systems.

Key Responsibilities

• Develop AI-Powered Decision Tools: Design, prototype, and implement agentic workflows and AI-assisted tools to automate complex safety evaluations and triage operational signals.
• Test & Validate Tool Reliability: Characterize the capabilities, edge cases, and failure modes of LLMs and AI workflows to ensure decision aids operate within safety-critical reliability thresholds.
• Integrate Data Pipelines: Connect generative AI agents to structured safety databases and operational logs for automated real-time synthesis and risk reporting.
• Collaborate Across Safety & Software: Work with safety strategists and software engineers to test prototypes in real operational scenarios and iterate based on user feedback.

Requirements & Tech Stack

• Currently pursuing a Master’s degree (or advanced degree) in Computer Science, Artificial Intelligence, Robotics, Data Science, or related engineering field.
• Good academic standing and enrolled at an accredited college or university throughout the internship.
• Expected graduation date between Fall 2027 and Summer 2028.
• Able to commit to a full-time, 12-week internship.
• Ability to relocate to the Bay Area, California, for the duration of the internship.
• Interns may not use proprietary information for thesis or publication outside of Zoox.

Technical Qualifications

• Strong software engineering fundamentals in Python and object-oriented programming.
• Hands-on experience developing and deploying AI tools, LLM APIs, and agentic workflows (e.g., LangChain, LlamaIndex, Claude/Cursor environments).
• Critical understanding of the limitations, hallucinations, and verification methods required when deploying AI in decision-making contexts.
• Experience building working prototypes, APIs, or internal web dashboards (e.g., Streamlit, FastAPI, or Flask).
• Strong analytical problem-solving and clear written/verbal technical communication.

Bonus Qualifications

• Prior experience applying AI/ML techniques to autonomous systems, safety analytics, or risk management.
• Experience designing evaluation harnesses to benchmark generative AI and prompt performance.
• Familiarity with database architectures (SQL/NoSQL) and data warehousing platforms.