AI and LLM Engineer (agents, tool use, MCP)

Role overview Build an AI assistant embedded in a floor planning editor, along with an MCP server that lets external AI applications work with plans through the same tools. The role focuses on making model-driven edits useful, measurable, and safe for users to review and undo.

Responsibilities - Design tools models use to inspect and edit plans, and build the execution loop around those tools. - Make each AI change reviewable and reversible, while enforcing user-defined limits. - Create evaluations that compare outcomes with saved plan data rather than judging only model responses. - Develop MCP server capabilities, including tools, resources, prompts, OAuth, and API keys. - Mitigate prompt injection and uncontrolled costs with budgets, confirmations, and rate limits.

Requirements - Production experience shipping LLM features that use tool or function calling. - Strong TypeScript skills, or experience in another typed language and willingness to work in TypeScript. - Experience evaluating model behavior and using evaluation results to improve systems. - Familiarity with streaming APIs such as server-sent events, retries, and timeouts.

Nice to have - Experience with MCP servers or clients, OAuth 2.1, image inputs such as sketches or plans, or spatial and geometric reasoning.

Benefits and work setup - Full-time engineering role, remote in India, the United States, the United Kingdom, Canada, Germany, the Netherlands, Poland, Australia, Singapore, or the United Arab Emirates.