Software Engineer, AI Application

The role

You will build the layer that lets agents participate in work by reading context from a workspace and taking action. This is product engineering involving writing Python and designing retrieval and tool surfaces.

What you will do

• Build agent workflows including planning loops, tool calling, multi-step execution, and recovery
• Design and ship MCP servers that expose Plane capabilities as tools
• Own the retrieval layer end to end including chunking, embedding, vector storage, and reranking
• Build agent interfaces including streaming output, intermediate step visibility, and approval gates
• Establish the evaluation harness defining datasets, regression suites, and metrics
• Instrument tracing, token cost, and latency across agent runs
• Design the permission and audit model for agent actions
• Evaluate models, frameworks, and techniques to make adoption recommendations
• Partner with backend, frontend, and product engineers to ship prototypes

Required qualifications

• Two or more years building production software
• At least one year shipping features backed by large language models
• Strong Python
• Experience operating services in production including debugging under real traffic
• Experience building agent systems that call tools
• Working knowledge of retrieval systems including embedding models, chunking, and vector databases
• Experience building an evaluation harness for a non-deterministic system
• Examples of AI features shipped to real users

Nice to have

• Built or published an MCP server
• Run agents against production data with a real permission model
• Worked on developer tools, project management, or another product with a complex domain model
• Contributed to open source

About the company

Plane is a work management platform that brings projects, knowledge, and requests into one open, adaptable platform. It is used by teams around the world and is built across San Francisco, London, and Hyderabad.