Skip to content

Reviewer Guide

Campus RAG Assistant is a production-style enterprise RAG platform for governed campus knowledge. It is intended to be reviewed through source code, architecture, screenshots, evaluation results, and operational artifacts — not as a hosted public product.

90-second read

  • Full-stack RAG platform built around governed institutional knowledge.
  • Multicloud provider boundaries across AWS Bedrock KB, Azure AI Search / OpenAI, and mock CI/local providers.
  • Vue 3 product UI with sessions, streaming chat, cited sources, feedback, OAuth handoff, and opt-in web research.
  • LangGraph orchestration, RAGAS evaluation baseline, LangSmith traces, Prometheus metrics, k6 load tests, CI/CD, and security docs.
  • Mock providers allow local and CI execution without cloud credentials.
  • Not presented as a hosted public product.

Senior signals

Signal Evidence
AI platform architecture Provider registry, AWS / Azure / mock separation, tenant config — ADR-001
RAG engineering LangGraph stages, multi-query retrieval, rerank hooks, explicit source contracts — LangGraph roadmap
Evaluation discipline RAGAS golden set, documented baseline, release-oriented gates — Evaluation
Observability LangSmith traces, request IDs, Prometheus metrics — Operations
Product judgment KB-first answers, opt-in web research, source transparency, feedback — Web Research roadmap
Production thinking CI/CD, gitleaks, dependency review, rate limits, load testing, hardening backlog — CI/CD, Security, Load Testing, Production Hardening

What this repository implements

This project builds from the public ets-berkeley-edu/chabot codebase and extends it into a source-reviewable AI platform. Implemented surface:

  • Vue 3 product UI (sessions, streaming, sources, feedback, OAuth handoff)
  • Provider registry for AWS, Azure, and mock execution modes
  • LangGraph RAG pipeline (condensemulti_queryretrievererankgenerateformat)
  • Tenant-hydrated prompt and config model in Postgres
  • RAGAS evaluation harness with documented baseline
  • LangSmith trace capture for KB and web research paths
  • Opt-in web research path with disclaimer UI and WEB-labeled sources
  • CI/CD, gitleaks, dependency review, k6 load testing, release docs, and a production hardening backlog

Suggested review paths

Reviewer Start here
Hiring manager Case Study
Staff / principal engineer Architecture, Design Notes, ADRs
AI engineer Evaluation, LangGraph roadmap, Provider registry ADR
Platform / DevOps reviewer CI/CD, Operations, Security, Load Testing
Product reviewer Screenshots, Case Study, Product Roadmap

What this is not

  • Not an official UC Berkeley or UC product.
  • Not a public hosted SaaS product.
  • Not a production deployment claim.
  • Not a generic chatbot demo.
  • Not a claim that the current evaluation set is production-sufficient.

See Notice and License for attribution and licensing details.