Full Stack Software Engineer, AI Integration

Ford Ford · Auto · Dearborn, MI +1 · Enterprise Technology

Full Stack Software Engineer focused on integrating AI, specifically LLMs, into applications. The role involves architecting systems with autonomous agents, utilizing a Model Context Protocol for live data interaction, and providing real-time streaming experiences. Key responsibilities include designing agentic loops, implementing tool-use capabilities, integrating with proprietary data silos, developing high-concurrency back-ends and streaming front-ends, optimizing RAG pipelines, and establishing AI evaluation and observability.

What you'd actually do

  1. Design Autonomous Loops: Transition from linear "chain" workflows to self-correcting agentic loops using frameworks like LangChain, LlamaIndex, or CrewAI.
  2. Tool-Use Architecture: Design and implement robust "tool-calling" capabilities, ensuring LLMs can reliably interact with external APIs and microservices.
  3. MCP Integration: Build and maintain Model Context Protocol (MCP) servers to bridge the gap between LLMs and our proprietary data silos securely and in real-time.
  4. High-Concurrency Back-end: Develop asynchronous Python (FastAPI) or Java services optimized for long-running AI tasks and token streaming.
  5. Streaming Front-end: Build responsive, stateful UIs in React or Angular that handle complex AI interactions (streaming text, generative UI components, and multi-modal feedback).

Skills

Required

  • BS in Computer Science or Engineering related field
  • 5+ years of total software engineering experience
  • 2+ years focused on AI integration
  • 2+ years of professional experience with LangChain, LlamaIndex, or Google ADK
  • GitHub Copilot (IDE & CLI), Claude Code, Cursor, or similar agentic coding assistants
  • Ability to architect end-to-end AI-native systems
  • 3+ years designing and building Python (FastAPI/Flask) or Java (Spring Boot) services
  • 2+ years designing real-time, streaming UI systems in React or Angular
  • Expert knowledge of SQL (PostgreSQL/pgvector) and Vector Databases (Chroma, Qdrant, or Pinecone)
  • Hands-on experience with GCP (Vertex AI) or AWS (Bedrock)
  • Experience with Git, Docker, Kubernetes,Terraform, and CI/CD (GitHub Actions/Jenkins)

Nice to have

  • The Agentic Mindset: You have a proven track record of moving models from "answering questions" to "completing multi-step tasks."
  • Prompt Engineering as Code: You treat prompts as production code—versioned, tested, and optimized for deterministic outcomes.
  • MCP Expertise: You understand the future of data grounding and have experimented with or implemented Model Context Protocol servers.
  • Growth Agility: You stay ahead of the curve, moving fluently from RAG-based architectures to Long-Context model strategies as the landscape shifts.

What the JD emphasized

  • AI-native applications
  • autonomous agents
  • agentic loops
  • tool-calling
  • streaming experiences
  • AI Evals
  • hallucination rates
  • latency
  • cost
  • automated AI benchmarking

Other signals

  • autonomous agents
  • LLMs as core engine
  • agentic loops
  • tool-use
  • streaming experiences