Gardner Diliman Assistant — Domain RAG System interface and evaluation output
Figure 01 — Gardner RAG AssistantOpen full resolution ↗

notesarchitecture & scope

A domain-grounded web assistant built for Gardner College Diliman. Combines retrieval-augmented generation (RAG) over curated institutional knowledge with Gemini language generation to answer student and visitor inquiries accurately while minimizing hallucination.

Key Capabilities

  • Context retrieval over structured institutional FAQs, academic programs, and policies
  • Streaming conversational responses powered by the Gemini API
  • Typed Next.js App Router backend and responsive client interface
  • Production deployment on Vercel with low-latency edge delivery

Demonstrates pragmatic RAG architecture: pairing narrow, verifiable context retrieval with a minimal chat interface.