RAG Main Quest: AI Memory Implant Scheme

Speaker

Vic Wen

Vic Wen is a lecturer at the NTUAI Club and a software engineer who specialises in building scalable infrastructure for AI. He also contributes to the open-source Apache Mahout project.

About the course

This lecture focuses on Retrieval-Augmented Generation, aka RAG. Starting from the limitations of large language models, it explains how retrieval mechanisms can enable AI to access external knowledge and build intelligent question-answering systems with practical value for business applications.

The content balances conceptual foundations with practical case studies, helping participants understand the core architecture of modern AI applications, rather than merely focusing on the use of prompts.

Teaching Topics

The course covers the following topics:

．Introduction of Retrieval-Augmented Generation (RAG)

．Limitations and Challenges of Large Language Models (LLMs)

．RAG System Architecture and Data Flow

．Chunking & Embedding

．Vector Databases and Retrieval Mechanisms

．Prompt Engineering 與 Context assembly

．Building a RAG System on Antigravity with Docker and Telegram Bot Integration

Technical Stack

The following technologies are covered in the course demonstration:

Retrieval-Augmented Generation (RAG)

Embedding Models

Vector Database

Semantic Search

Prompt Engineering

Python

Docker

Git

Telegram Bot