Knowledge-base search in production backends

Enterprise RAG & Conversational Agents

Status: In production
Role: AI Engineer
Organisation: Robionix Technologies
Year: 2024-2026

The problem

Company knowledge sits in documents nobody can find, and a chatbot that hallucinates the answer is worse than no chatbot.

Outcome

Integration: REST into Laravel + ReactIntegration
Retrieval: Vector-backedRetrievalFAISS / Chroma.

A policy question answered with the exact source chunks cited alongside — grounded, not guessed. — The Nexus RAG chatbot answering a company-policy question with FAISS retrieval results and source citations shown alongside.

The Nexus RAG chatbot answering a company-policy question with FAISS retrieval results and source citations shown alongside.

Approach

01
Built RAG pipelines over vector databases with LangChain and LangGraph, driven by the Claude and OpenAI APIs.
02
Served them as inference REST APIs consumed by existing Laravel, React and MySQL production backends - integration into a live stack, not a standalone demo.

Stack

LangChain
LangGraph
Claude API
OpenAI API
FAISS
Chroma
FastAPI
MySQL