Back to Case StudiesView on GitHub
OCI
RAG-Based HLD Document Generation System
Architected a production RAG system on OCI Generative AI that automates High-Level Design document generation — cutting generation time by 50-70% while controlling hallucinations via grounded prompting.
OCIAI/MLLangChainOracle DB 23ai
The Problem
Enterprise delivery teams were spending weeks manually authoring HLD documents for each client engagement, leading to inconsistency and delays in project kick-off.
The Solution
Built a RAG pipeline using OCI Gen AI (LLM inference), Oracle DB 23ai vector search for knowledge base retrieval, LangChain for orchestration, and a Next.js + FastAPI production stack for the interface.
Results
40%Faster KB retrieval latency
50-70%Reduction in HLD generation time
~0Hallucination rate with grounded prompting
Tech Stack
PythonLangChainLangGraphOCI Gen AIOracle DB 23aiNext.jsFastAPI