LLM-RAG Agent Integration

Budget: ₹5000 - ₹10000

I’m looking for someone who can take separate language models, agent-based workflows, and Retrieval-Augmented Generation components and weave them into a single, reliable stack. The core goal is integration—getting these pieces to talk to each other smoothly, surface accurate context on demand, and expose a clean API that downstream apps can call without worrying about the plumbing.

What I already have: standalone LLM endpoints, a small library of task-specific agents, and an indexed knowledge base ready for RAG queries. What I need from you: the connective tissue. That means orchestrating calls between the models and agents, handling vector or hybrid search for the RAG layer, and implementing guardrails so the final responses stay on topic and safe.

Deliverables • End-to-end workflow that links the LLM, agent manager, and RAG retriever • Clear, commented code or notebooks showing how each call is made • Minimal API (REST or FastAPI is fine) that external services can hit with a prompt and receive a consolidated answer • Read-me style documentation covering setup, environment variables, and extension points

Acceptance criteria • Given a test prompt, the system retrieves relevant context, routes it through the correct agent, and returns a coherent response in under two seconds. • All major components launch with one command in a fresh environment. • Unit tests demonstrate at least 90 % coverage of the integration logic.

Tools are flexible