We are seeking a Senior AI Engineer to join an enterprise AI platform engineering initiative delivering a Databricks-native, MCP-first platform with a Hub & Spoke governance model, enabling independent spoke teams to build and deploy AI agents across GCP and Azure environments. In this role, you will design and implement core AI agents on the platform, from specification through production-ready deployment, ensuring accurate, reliable, and fully integrated solutions within the platform's shared services. Responsibilities Participate in use case deep-dive sessions and translate business requirements into specifications using a 15-characteristic framework covering scope, tools, memory, guardrails, evaluation criteria, and acceptance definition Design and implement the RAG pipeline, including document retrieval, chunking strategy, embedding, reranking, and response generation Integrate agents with platform MCP servers for tool access and with the LLM Gateway for model routing Implement prompt engineering practices such as system prompts, few-shot examples, and chain-of-thought patterns, and iterate based on evaluation results Wire guardrails, including prompt injection protection and domain boundary enforcement, for reference agents Run evaluation cycles using the platform Evaluation Framework, including LLM-as-judge scoring and RAG faithfulness and relevance metrics, and iterate until quality gate criteria are met Collaborate with the Data Engineer on data schema and retrieval interface design, and with the QA Engineer on test coverage and the acceptance test query set Support UAT with business stakeholders, address feedback, and prepare agents for production deployment following the platform runbook Requirements 3+ years of experience building LLM-based applications in Python Hands-on experience with RAG pipelines, covering retrieval, chunking, embedding, reranking, and generation Proficiency in prompt engineering, including system prompt design, few-shot patterns, and output structuring Experience with LLM orchestration frameworks such as LangChain, LlamaIndex, or equivalent Familiarity with LLM evaluation techniques, including LLM-as-judge, RAG faithfulness metrics, and RAGAS or equivalent Expertise in Databricks and MLflow, including experiment tracking and model serving awareness Knowledge of MCP (Model Context Protocol) or equivalent tool-use / function-calling patterns Experience with vector stores such as Chroma, Pinecone, or Databricks Vector Search, or equivalent Skills in REST API integration and OAuth authentication flows Strong product thinking, with the ability to connect technical implementation choices to user-facing quality outcomes Iterative mindset, comfortable with repeated evaluation-tune-evaluate cycles without losing focus on delivery deadlines Clear communication with non-technical business stakeholders during UAT and discovery sessions Proficiency in English at a B2+ level Nice to have Prior delivery of a production RAG or agentic AI system end-to-end Experience working within a governed AI platform, such as an LLM gateway, guardrails, and evaluation framework, rather than fully custom stacks Familiarity with responsible AI concepts, including hallucination, grounding, PII, and prompt injection