What project we have for you

The AI Product Family delivers enterprise LLM agents and AI services integrated with our Microsoft ecosystem. We run on Azure (Azure OpenAI/LLMs, Azure Functions, AI Search/vector, Service Bus, Storage, Key Vault) and connect securely to CRM, ERP, Microsoft Graph, and SharePoint. You’ll establish patterns for prompt engineering, multi-agent design, RAG, and ship reusable components that accelerate future solutions.

What you will do

• Architect, develop, and optimize LLM-powered agents using frameworks such as Agno AI, Pydantic AI, LangChain, LangGraph, and LlamaIndex.
• Implement retrieval-augmented generation (RAG) workflows, integrating vector databases, embeddings, and semantic search capabilities.
• Deploy, fine-tune, and evaluate LLMs (OpenAI, Azure OpenAI, Claude, LLaMA, and others) for a range of business use cases.
• Integrate AI agents with external APIs, internal platforms, and business data sources.
• Collaborate with IT engineers and business stakeholders to define requirements and deliver robust, secure AI-enabled MVPs.
• Mentor internal teams on AI/ML best practices, tooling, and prompt engineering.

What you need for this

Must have

• Prompting and proven experience with LLMs such as OpenAI, Azure OpenAI, Claude, and LLaMA.
• Strong hands-on expertise in agentic orchestration frameworks, such as (any) Agno AI, Pydantic AI, LangChain & LangGraph.
• Practical knowledge of vector databases (e.g., Pinecone, Azure Cognitive Search), and RAG and its techniques in general.
• Solid programming skills in Python.
• Experience integrating AI solutions with external APIs and internal platforms.
• Understanding of agent protocols such as MCP and A2A.
• Excellent problem-solving skills, autonomy, and a collaborative mindset.

Will be a plus

• Familiarity with .NET frameworks such as Microsoft 365 Agents SDK (Bot Framework SDK) and Semantic Kernel.
• Experience with Azure AI services, ideally Azure AI foundry.
• Ability to mentor teams on prompt engineering and AI/ML best practices.