Case Study
EduSathi
Production-grade AI study companion utilizing RAG, FAISS, LangChain, and Groq API with dual-portal RBAC for university curricula.
What is EduSathi?
EduSathi is an AI-powered study assistant designed specifically for university students, leveraging a 4-stage Retrieval-Augmented Generation (RAG) pipeline to deliver precise, curriculum-aligned answers.
Problem
Students often struggle to find specific, reliable answers within hundreds of pages of university-provided PDF materials. Traditional keyword searches are inefficient, and generic LLMs hallucinate or provide out-of-syllabus answers.
Architecture
Ingestion: High-speed PDF parsing
Chunking: Recursive text boundaries
Indexing: FAISS dense vector search
Inference: Groq LLaMA models
Tech Stack
PythonLangChainFAISSGroq APIStreamlit
Architecture Benchmark • Comparative Analysis
| Capability | Traditional Search | Generic LLM | EduSathi RAG Pipeline |
|---|---|---|---|
| Context Grounding | Keyword matches only | High hallucination risk | Strictly grounded in syllabus PDFs |
| Vector Search | None | Static training weights | FAISS dense semantic indexing |
| Inference Speed | Instant (no synthesis) | 2–5s (Standard APIs) | Sub-second Groq LLaMA 3.1 |
| Governance | None | Unrestricted | Dual-portal RBAC (Student/Faculty) |