Mcp/ai Developer

Autodesk Autodesk · Enterprise · Toronto, ON +1 · Remote

Autodesk is building an agentic AI platform for their Viewer MCP, enabling users to interact with models in natural language. This role involves building end-to-end features including MCP tools, agentic workflows, and supporting services, integrating LLMs and models via APIs, and developing guardrails for agentic workflows.

What you'd actually do

  1. Implement and maintain clean, well-tested code for MCP tools, agentic features, and their supporting services (front-end and back-end), to team standards.
  2. Integrate LLMs and models via APIs/SDKs to deliver agentic workflows, with attention to correctness, latency, and cost.
  3. Help build the guardrails of agentic workflows — human-in-the-loop steps, traceability, and clear error handling.
  4. Contribute to automated tests, evaluation harnesses for AI behavior, CI/CD pipelines, and developer tooling.
  5. Debug and troubleshoot across the stack (agent/tool logic, APIs, web UI), collaborating with senior engineers and QA.

Skills

Required

  • Master's in Computer Science / Computer Engineering, or a Bachelor's with 3+ years of relevant experience, or equivalent practical experience.
  • Hands-on experience building software in at least one mainstream language (TypeScript/JavaScript preferred for our stack).
  • Familiarity with modern web development (HTML/CSS/JS) and a framework such as React.
  • Sound understanding of testing, debugging, and code review; experience with Git and collaborative workflows (PRs, branches).
  • Exposure to building with LLMs or AI APIs — e.g. prompting, calling model/agent APIs, or simple agentic/RAG features (course, side-project, or work experience all count).
  • Strong problem-solving skills, eagerness to learn ML/AI deeply, and a collaborative mindset.
  • Experience with Agile practices; strong written and spoken English.

Nice to have

  • Node.js back-end development and RESTful APIs.
  • Familiarity with the Model Context Protocol (MCP), agent frameworks, or tool-calling patterns.
  • Cloud and CI/CD basics (AWS, Docker, GitHub Actions) and observability tooling.
  • Foundational ML knowledge (embeddings, vector search/RAG, model evaluation).
  • Exposure to 2D/3D rendering, visualization, or graphics concepts (WebGL/WebGPU).

What the JD emphasized

  • agentic AI
  • agentic platform
  • agentic workflows

Other signals

  • agentic AI
  • agentic platform
  • agentic workflows
  • integrating LLMs and models