Principal Full Stack AI Engineer (LLMs, Agentic AI & React/Typescript)

Job Requisition ID #

26WD97978

Position Overview

The DMD team builds AI agents that help Autodesk users work with design and manufacturing data access and collaboration. You'll contribute across the full stack - backend agent logic and the React/TypeScript UIs that surface it to users.

Responsibilities

• Lead and deliver full-stack features across backend agent services and the React/TypeScript UI, from architecture through production operation.
• Drive technical decisions on system architecture, agent behavior, tooling, and developer experience
• Own production reliability, security, and observability
• Mentor teammates and set quality standards across the codebase, including testing, security, and operational rigor.
• Own features end-to-end from design through deployment and ongoing operation.

• 10-14 years of software engineering experience, with a track record of delivering large-scale production systems
• Hands-on experience with LLM application patterns: tool use, streaming responses, multi-turn conversation, and context/session management
• Production experience with at least one agent framework (e.g. Strands Agents, LangGraph, AutoGen) and/or MCP (Model Context Protocol) or A2A (Agent-to-Agent) protocols
• Proficient in Python/Java - async patterns, testing, clean module design
• Experience building and deploying REST/streaming APIs on AWS (Lambda, API Gateway, IAM, Secrets Manager)
• Production experience with React and TypeScript
• Experience hardening AI/agent systems for production: authorization boundaries, dependency resilience (retries, timeouts, circuit breakers), and observability (metrics, tracing, structured logging)
• Comfortable leading in a polyglot monorepo
• Actively uses AI coding tools (e.g. Claude Code, Cursor, Copilot) as part of daily workflow
• Strong communication skills - able to drive trade-off discussions and work autonomously in a small team

• Familiarity with RAG pipelines or vector search (embeddings, semantic similarity thresholds)
• Prompt engineering / system prompt design experience
• Experience with spec-driven development workflows (e.g. OpenSpec)