Key responsibilities :

1. Use-case identification and prioritization :

- Partner with business and engineering leaders to build and rank a portfolio of GenAI use cases by impact, feasibility, and risk; help maintain the value backlog reviewed monthly by the joint governance committee.

2. Shared architecture and patterns :

- Work with architects and engineers across global teams to shape a shared blueprint for GenAI on Azure - Azure AI Foundry, Azure OpenAI, Azure AI Search, and the Microsoft Agent Framework - and co-develop reusable patterns for RAG, agents, evaluation, and guardrails that every embedded pod can adopt.

3. Hands-on solution design :

- Co-design and prototype flagship use cases end to end with the engineering teams : retrieval, orchestration, model selection, grounding, and integration with enterprise systems and data.

4. Responsible AI and governance :

- Embed enterprise AI governance and model-risk management - evaluations, content safety, access control, and auditability - with awareness of NIST AI RMF and the EU AI Act where applicable, for regulated and sensitive data.

5. Mentoring and enablement :

- Support and mentor implementation engineers, facilitate design reviews, and help raise GenAI maturity across the teams you work with, advancing them along the LLMOps maturity model.

6. Outcomes and impact :

- Help instrument and report productivity, time-to-market, and quality gains against the program's shared targets.

Must-have qualifications :

- 12+ years in software / ML engineering, with 3+ years shipping production GenAI/LLM systems, as architect or principal engineer.

- Deep Azure AI : Azure AI Foundry, Azure OpenAI (GPT-5 family incl. GPT-5.5, plus o-series reasoning), Azure AI Search (vector + hybrid + semantic ranking), Azure AI Document Intelligence (intelligent document processing), Azure ML and Prompt Flow.

- Agentic systems : multi-agent orchestration with the Microsoft Agent Framework (Semantic Kernel / AutoGen) and LangChain / LangGraph; tool and function calling.

- Enterprise RAG : chunking, embeddings, retrieval tuning, reranking, grounding, and RAG evaluation.

- Data platforms : Snowflake (incl. Cortex AI - Cortex Search / LLM functions), Microsoft Fabric / Azure Databricks, for grounding, features, and analytics.

- Expert Python (incl. async); prompt engineering as versioned, tested, reviewed code.

- LLMOps : evaluation harnesses, observability, cost / latency optimization, CI/CD (Azure DevOps / GitHub), AKS, Azure Functions / App Service.

- Responsible AI : Azure AI Content Safety, guardrails, prompt injection / jailbreak defense, document-level access control for regulated data.