Role Overview
The AI Product Developer – Full Stack designs, builds, and operates scalable AI‑enabled web applications, integrating large language models, retrieval‑augmented generation, and agent frameworks on the Azure platform.
Responsibilities
- Own end‑to‑end delivery of AI agent applications from requirement to production.
- Build full‑stack applications using React, Angular or Vue with Python back‑end services and RESTful APIs.
- Design and implement RAG pipelines: document ingestion, parsing, chunking, embeddings, vector indexing, hybrid search, reranking, grounding, and response generation.
- Develop LLM‑powered agents with tool calling, memory, workflow orchestration, human‑in‑the‑loop controls, and multi‑agent patterns.
- Integrate agents with enterprise systems, databases, APIs, and enforce security, authentication, secrets management, privacy, and responsible AI controls.
- Establish evaluation, observability, and testing for LLM applications, including relevance, groundedness, latency, token usage, cost, safety, and regression testing.
- Implement CI/CD, containerization, monitoring, and Azure deployment using GitHub Actions and Azure services.
Requirements
- 5‑7 years of professional Python development with FAST API/ASP.NET, object‑oriented design, asynchronous programming, and production‑grade REST APIs.
- Hands‑on full‑stack experience with React, Angular or Vue, modern JavaScript/TypeScript, component‑driven UI, state management, API integration, authentication, and responsive design.
- Strong SQL and database design skills; experience with relational, NoSQL, document, and vector data stores.
- Proven experience deploying and operating applications on Azure, including hosting, identity and access management, networking, monitoring, security, secrets management, and service integration.
- Experience building RAG pipelines with document loaders, parsing, chunking, embeddings, vector databases, metadata filtering, hybrid retrieval, reranking, citation, and context assembly.
- Solid understanding of LLM patterns such as prompt engineering, tool and function calling, structured outputs, conversation memory, context‑window management, model routing, caching, and workflow orchestration.
- Experience with AI agent frameworks (Semantic Kernel, LangChain, LangGraph) for agent state, tool execution, guardrails, human approval, and multi‑agent coordination.