Objective
This is a 16 weeks training for LLM engineer. It’s designed to cover both theoretical foundations and practical implementations, allowing participants to gain hands-on experience in working with LLMs. Among our 2024 students, 80% of graduates found relevant high-paying jobs.
这是针对 LLM 工程师的 16 周培训。它旨在涵盖理论基础和实际操作，让参与者获得使用 LLM 的实践经验。在我们2024的学生当中，有80%的毕业生找到了相关的高薪工作。
Syllabus:
Version-2(2025)
| Model | content | hours |
| 1 | Introduction to Large Language Models | 12 |
| 2 | LLM technologies | 20 |
| 3 | LLM app solution and platform | 20 |
| 4 | Advanced topics | 4 |
| 5 | Final Project and Evaluation | 8 |
Module 1
- Introduction to Large Language Models (12 hours)
- AI Basic
- Deep learning for LLM
- NLP foundation for LLM
- Transformer Architecture and Core Concepts
- Pre-Training and Fine-Tuning
- LLM Training Pipeline
- LLM Deployment and Optimization
Module 2
- LLM technologies ( 20 hours)
- Prompt engineering
- RAG system and its optimization
- Coding lab
- Fine tuning ( 2 hours)
- Coding lab
- Agent system ( 2 hours)
- Coding lab
Module 3
- LLM app solutions platforms (20 hours)
- Server client architect / SAAS
- OpenAI APIs
- Opensource LLM model
- Langchain
- Hugging Face LLM
Module 4
- Advanced topics (4 hours)
- Hallucination and how to solve
- LLM Ethics, Bias
- Security / safety
Module 5
- Final Project and Evaluation
- Project name: LawGPT
The Law GPT application leverages advanced AI technologies to offer a user-friendly platform for legal assistance, focusing particularly on immigration law. It integrates LLM functions and RAG (Retrieval-Augmented Generation) to enhance response quality by grounding answers in relevant documents. The system handles diverse user queries by generating insightful, context-aware responses.
下一期开始日期：
- Sat, Mar 8th, 2025
- 上课时间：
每个周六，周三，5-7PM美西时间，9-11AM北京时间
