Sage AI

SAGE AI — Smart AI Guide for Education

Next.js 15, Supabase (PostgreSQL + pgvector), Ollama (Local LLM), RAG

Built a full-stack, RAG-based academic study assistant enabling students to query their own course materials and receive

grounded, source-cited answers from a private knowledge base.

Engineered a semantic retrieval pipeline with vector embeddings in pgvector, running LLMs fully on-device to eliminate API

costs and keep student data private.

Delivered as a 6-member team under academic supervision, owning the backend and AI-integration layers end-to-end.