About the role
This role builds the core agentic infrastructure orchestration engine, RAG pipelines, prompt templates, API layer, and self-learning feedback loops for a reusable, cross-organization agentic AI platform.
Responsibilities:
- Build the orchestration engine for agentic workflows, including prompt template implementation, tool-calling, and dynamic routing.
- Develop and maintain the RAG engine embedding, retrieval, and context injection across multiple data sources.
- Implement feedback loops and self-learning pipelines for continuous accuracy improvement.
- Develop REST APIs to connect orchestration layers to model endpoints and downstream services.
- Build NL-to-SQL pipeline components within the agentic platform.
Requirements:
- Good to have, maybe hard to find combined: Feedback loop and self-learning pipeline implementation for continuous model improvement.
Must-Have:
- Python development, REST API development.
- LangChain or LlamaIndex orchestration framework implementation.
- Azure OpenAI integration and consumption.
- Redis session and context management within agentic workflows.
- Vector database implementation and integration (e. g., Pinecone, Azure AI Search, Qdrant).
- NL-to-SQL technique production implementation within agentic pipelines.
- RAG engine design and building embedding, retrieval, and context injection pipelines.