Remote - Senior AI Engineer

Role Summary
This role builds the conversational analytics and agent layer on top of the Redshift presentation data. You will wire enterprise data into Microsoft Copilot via Azure AI Foundry, develop and deploy AI agents consumed across Microsoft 365, and build the RAG pipeline for the legacy unstructured data use case. Context mapping is a core function — you will document and structure data semantics so agents retrieve accurate, relevant answers. Strong Python, LLM API integration, and a data engineering foundation are essential.
Key Responsibilities
Connect Amazon Redshift data to Microsoft Copilot via Azure AI Foundry
Build, test, and deploy AI agents for consumption in email, Word, Excel, PowerPoint, and Copilot
Design and maintain a portable semantic layer in YAML files versioned in Git
Map business context to data entities to enable accurate agent retrieval
Optimise MCP token usage for enterprise-scale agent deployment
Build a RAG pipeline over legacy unstructured factory floor data
Normalise and embed unstructured documents (PDFs, specs, photos metadata) into a vector store
Build retrieval agents for historical manufacturing case search
Collaborate with the Senior DE to ensure clean data surfaces for agent consumption
Document agent logic, context mappings, and retrieval schemas
Requirements
6+ years in software or data engineering with at least 12 months in AI/LLM development
Strong Python — comfortable building production agent pipelines
Hands-on experience with Azure AI Foundry or a comparable agent orchestration platform (AWS Bedrock, LangChain, etc.)
Solid understanding of RAG architecture, vector databases, and document intelligence
Experience with Microsoft Copilot extensibility or Microsoft 365 developer integrations
Ability to perform context mapping: translate business questions into data retrieval logic
Familiarity with semantic layer concepts and YAML-based metadata management
Understanding of token optimisation and prompt engineering for enterprise deployments
Nice to Have
Exposure to Microsoft Fabric, Power BI semantic models, or Dataverse
Experience with unstructured data normalisation and document intelligence pipelines
Familiarity with Amazon Redshift or AWS data infrastructure
Prior engagement in a managed services or consulting delivery model
Interest in agentic data engineering — using agents to generate transformation logic from contracts

Work arrangement
Yes