TLEF ENGE AI

EngE-AI is an AI-powered learning assistant for UBC Engineering courses which emphasizes on developing student's critical thinking.

Features

Instructor

• Course creation and management
• Document upload (PDF, DOCX) for RAG
• Learning objectives and materials
• Flags management and responses
• Student chat monitoring
• Assistant and system prompts configuration

Student

• Course-aware AI chat
• Flag creation ("I'm struggling")
• Flag history
• Access to course materials and objectives

Technical

• Vector search (Qdrant)
• Streaming chat responses
• Session-based authentication
• MongoDB for courses and users
• Custom UBC Shibboleth SAML 2.0 authentication strategy for Passport.js

---

Recent Update

---

How to Set It Up

Prerequisites

• Node.js (v18+)
• MongoDB (running and accessible)
• Qdrant (vector database)
• LLM endpoint (e.g. Ollama or other provider)
• SAML (optional): IdP metadata, issuer, callback URL for CWL auth. For local development, use docker-simple-saml as a containerized IdP.

Setup Steps

1. Clone the repo:

   git clone https://github.com/ubc/tlef-engeai.git
   cd tlef-engeai

2. Install dependencies:

   npm install

3. Create a .env file in the project root (see Environment Variables below).

4. Run the application:

   • Development: npm run dev (nodemon + BrowserSync)
   • Production: npm start

Environment Variables

Group	Variables
Server	TLEF_ENGE_AI_PORT (default 8020)
MongoDB	MONGO_HOST, MONGO_PORT, MONGO_USERNAME, MONGO_PASSWORD, MONGO_AUTH_SOURCE, MONGO_DB_NAME
Qdrant	QDRANT_URL, QDRANT_COLLECTION_NAME, QDRANT_VECTOR_SIZE, QDRANT_DISTANCE_METRIC, QDRANT_API_KEY (optional)
LLM	LLM_PROVIDER, LLM_ENDPOINT, LLM_DEFAULT_MODEL, LLM_API_KEY
Embeddings	EMBEDDING_PROVIDER, EMBEDDINGS_ENDPOINT, EMBEDDINGS_MODEL
RAG	RAG_CHUNK_SIZE, RAG_OVERLAP_SIZE, RAG_CHUNKING_STRATEGY, RAG_MIN_CHUNK_SIZE
SAML	SAML_AVAILABLE, SAML_ISSUER, SAML_CALLBACK_URL, SAML_ENTRY_POINT, SAML_LOGOUT_URL, SAML_METADATA_URL, SAML_ENVIRONMENT
Session	SESSION_SECRET, SESSION_TIMEOUT_MS
Optional	DEBUG, DEVELOPING_MODE (mock LLM), instructor PUID overrides

---

Teams

Role	Name
Principal Investigator	Alireza Bagherzadeh — Associate Professor of Teaching, CHBE
Co-Investigator	Amir M. Dehkhoda — Assistant Professor of Teaching, Materials Engineering
Software Developer	Richard Tape
Software Developer	Charisma Rusdiyanto

---

How to Contribute

1. Fork the repo
2. Create a feature branch
3. Follow existing code style (TypeScript, Express patterns)
4. Submit a pull request

For API reference, see documentation/ENDPOINT_ARCHITECTURE.md.

---

Documentation

• Endpoint Architecture
• Responsive Design
• docker-simple-saml — Containerized SAML 2.0 IdP for local development

---

Continuous Integration

Pushing to the main branch in this repo will trigger a deploy automatically to the staging server.