Full Stack Engineer, Backend Oriented (Accelerator Program, Intern)

About the role

An internship on an AI product team, working on the agent runtime and AI infrastructure that powers AI-native experiences. Interns help move ideas from prototype through benchmarking into production.

What you will do

• Improve core agent runtime pieces such as intent routing, query rewriting, RAG and multi-step workflows
• Design and tune tool calling features, covering discovery, execution, result handling and service integration
• Create evaluation datasets and harnesses that measure routing accuracy, answer quality and agent reliability
• Analyze failures, spot regressions and refine prompts, workflows and guardrails to raise agent quality
• Add tracing, metrics, logging and dashboards to strengthen observability and production monitoring
• Debug asynchronous services and agent pipelines, and write integration tests with mocked LLM responses

What they look for

• Solid software engineering grounding with strong Python or Java coding skills
• Good grasp of algorithms, data structures and problem solving, including system trade-offs
• Practical exposure to LLM applications or agent systems via research, coursework, internships or projects
• Strong debugging and analytical skills, with curiosity about why something fails

Nice to have

• Exposure to MCP, tool ecosystems or agent runtime work
• Background building benchmark, evaluation or LLM testing infrastructure