Course Name: Advanced Generative AI: End to end development with LLMs
51 Hours
Understand Generative AI (GAI) capabilities & assess business use cases based on latest GAI developments
Understand at high level how GAI assistant works & perform performance improvement on a stock LLM chatbot
Configure and deploy a LLM-powered chatbot with your personal touch / selected domain
Understand Generative AI (GAI) capabilities & assess business use cases based on latest GAI developments
Understand at high level how GAI assistant works & perform performance improvement on a stock LLM chatbot
Configure and deploy a LLM-powered chatbot with your personal touch / selected domain
Generative AI: End-to-end development with LLMs
Course Introduction
Generative AI: End-to-end development with LLMs. This comprehensive course is designed to teach you how to quickly harness the power of the Large Language Model (LLM).
From understanding what LLMs are, how it works to putting them to use and optimize through prompt engineering, model orchestration, frontend development to performance tuning. The course will equip you with the skills and knowledge necessary to develop cutting-edge LLM solutions for selected domains.
Course Contents
• Understand Generative AI (GAI) capabilities
• Assess business use cases based on latest GAI developments
• Understand at high level how GAI assistant works
• Performance tuning on a stock LLM chatbot
• Configure and deploy a LLM-powered chatbot with your personal touch / selected domain
Trainer
Henry Fong
Experience
Manage generative AI initiative for global banks and insurances in technology consultancy.
Strong technical background from data science, full stack development, cloud engineering, etc.
• Msc. in Data Analytics,
• B. Eng in Software Engineering
• AWS and GCP certified.
Understanding what neuron network is and how to build it with PyTorch
Content: Structure of neuron network. Why does it matter? Build your first NN using PyTorch
WEEK 5 Understanding Neuron Networks: 2/3
More on PyTorch. Why is GAN important?
Content: More hands-on development with Pytorch (e.g. CNN). Introduction to GAN.
WEEK 6 Understanding Neuron Networks: 3/3
Have fun on building your own GPT model
Content: Build your own GPT model with PyTorch (nanoGPT)
Module 3 ┊ Introduction to Cloud
WEEK 7 Introduction to Cloud: 1/3
Introduction to most common cloud services. Introduction to docker.
Content: Compute, block storage, serverless, VPCs. Docker, Jupyter in docker, build your first docker image.
WEEK 8 Introduction to Cloud: 2/3
Introduction to kubernetes. Running your first Kubernetes Cluster and introduction to Terraform.
Content: Introduction to Kubernetes. Running your first cluster (with GPU). Replicate the cluster with Terraform script.
Module 4 ┊ Introduction to LLM
WEEK 9 Introduction to LLM: 1/3
Play around with LLMs on the market and understand the difference. Deploy your first LLM.
Content: Understanding LLMs configurations and parameters. Deploy a minimum LLM virtual assistant.
WEEK 10 Introduction to LLM: 2/3
Natural Language Processing: Why LLM rocks (transformer). Introduction to Prompt Engineering.
Content: Introduction to legacy NLP. What is transformer. Basic prompt engineering technical and concepts.
WEEK 11 Introduction to LLM: 3/3
Common LLM problems. How to improve LLM’s performance: PEFT and RAG
Content: Introduction to common LLM issues such as hallucination. Model fine-tuning vs RAG.
Module 5 ┊ LLM Orchestration
WEEK 12 LLM Orchestration: 1/3
What is LLM Orchestration and why do we need it. Introduction to LangChain.
Content: Introduction to LangChain concepts.
WEEK 13 LLM Orchestration: 2/3
Building RAG using Langchain. What is embedding? What is vector database? What is similarity search?
Content: Hands-on session on building a RAG using Langchain and vector database. Address common embedding issues.
WEEK 14 LLM Orchestration: 3/3
Dealing with structured database with Langchain. How to write your pre/post processor? Connecting your LLM to GUI.
Content: Hands-on session on using Langchain and structured data. How to extend Langchain. Building a simple GUI for your virtual assistant (e.g. streamlit).
Module 6 ┊ Capstone Project
WEEK 15 Capstone Project: 1/4
Picking your problem statement. Defining deliverables and acceptance criteria. Prepare solution architecture, model selection
WEEK 16 Capstone Project: 2/4
Preparing data, decide how to improve the performance (fine-tuning vs prompt engineering). Data pipeline design and implementation.
WEEK 17 Capstone Project: 3/4
Baseline your LLM’s performance, improve prompt template. Resolve known issues.
WEEK 18 Capstone Project: 4/4
Prepare terraform for the solution. Deploying to Cloud. Troubleshooting.
Understanding what neuron network is and how to build it with PyTorch
Content: Structure of neuron network. Why does it matter? Build your first NN using PyTorch
WEEK 5 Understanding Neuron Networks: 2/3
More on PyTorch. Why is GAN important?
Content: More hands-on development with Pytorch (e.g. CNN). Introduction to GAN.
WEEK 6 Understanding Neuron Networks: 3/3
Have fun on building your own GPT model
Content: Build your own GPT model with PyTorch (nanoGPT)
Module 3 ┊ Introduction to Cloud
WEEK 7 Introduction to Cloud: 1/3
Introduction to most common cloud services. Introduction to docker.
Content: Compute, block storage, serverless, VPCs. Docker, Jupyter in docker, build your first docker image.
WEEK 8 Introduction to Cloud: 2/3
Introduction to kubernetes. Running your first Kubernetes Cluster and introduction to Terraform.
Content: Introduction to Kubernetes. Running your first cluster (with GPU). Replicate the cluster with Terraform script.
Module 4 ┊ Introduction to LLM
WEEK 9 Introduction to LLM: 1/3
Play around with LLMs on the market and understand the difference. Deploy your first LLM.
Content: Understanding LLMs configurations and parameters. Deploy a minimum LLM virtual assistant.
WEEK 10 Introduction to LLM: 2/3
Natural Language Processing: Why LLM rocks (transformer). Introduction to Prompt Engineering.
Content: Introduction to legacy NLP. What is transformer. Basic prompt engineering technical and concepts.
WEEK 11 Introduction to LLM: 3/3
Common LLM problems. How to improve LLM’s performance: PEFT and RAG
Content: Introduction to common LLM issues such as hallucination. Model fine-tuning vs RAG.
Module 5 ┊ LLM Orchestration
WEEK 12 LLM Orchestration: 1/3
What is LLM Orchestration and why do we need it. Introduction to LangChain.
Content: Introduction to LangChain concepts.
WEEK 13 LLM Orchestration: 2/3
Building RAG using Langchain. What is embedding? What is vector database? What is similarity search?
Content: Hands-on session on building a RAG using Langchain and vector database. Address common embedding issues.
WEEK 14 LLM Orchestration: 3/3
Dealing with structured database with Langchain. How to write your pre/post processor? Connecting your LLM to GUI.
Content: Hands-on session on using Langchain and structured data. How to extend Langchain. Building a simple GUI for your virtual assistant (e.g. streamlit).
Module 6 ┊ Capstone Project
WEEK 15 Capstone Project: 1/4
Picking your problem statement. Defining deliverables and acceptance criteria. Prepare solution architecture, model selection
WEEK 16 Capstone Project: 2/4
Preparing data, decide how to improve the performance (fine-tuning vs prompt engineering). Data pipeline design and implementation.
WEEK 17 Capstone Project: 3/4
Baseline your LLM’s performance, improve prompt template. Resolve known issues.
WEEK 18 Capstone Project: 4/4
Prepare terraform for the solution. Deploying to Cloud. Troubleshooting.
Upcoming Class
30 May 2025
- 51 Hours
- Every Friday 7:00pm-10:00pm
- Diamond Hill Classroom
Free Seminar
Course Seminar
- Starts at 7:30PM
- 2 Hours
- Hosted in our office at Causeway Bay
A1: To enroll in this course, you should have a general understanding of IT concepts, such as the difference between front-end and back-end development. A basic familiarity with programming concepts like loops and for loops is also recommended. While knowledge of Python is a significant advantage, it is not a strict requirement for this course.
A2: Studying this course is beneficial even if you're not planning a career in IT. The skill of developing generative AI use-cases is rare and highly valued in the market. By understanding how to integrate AI into business, you can significantly enhance your value in various job markets. The course also prepares you for the upcoming trend of innovation within various business sectors.
A3: This course is highly suitable for those considering a career switch to IT. AI's presence in Hong Kong is expanding, and large companies are likely to establish AI and workplace automation teams in the near future. By taking this course, you can practice these skills in your current job, gaining relevant experience for your transition into the IT industry.
A4: After completing the course, you can explore a variety of job opportunities. Roles such as AI Specialist and Business Analyst are particularly relevant, as they involve a deep understanding of AI applications in business.
It means accounting team, procurement team, HR team, Sales team, marketing team, are all looking for someone with their domain knowledge, to also know Generative AI to facilitate innovation within the team.
This will be the trend moving forward.
After studying this course, you may completely switch field to IT with the positions mentioned above, or you may seek to stay in your current industry but look for a new role that encourages innovation within your team.
A5: If you aspire to become a Business Analyst, studying a course in generative AI is highly beneficial as it equips you with cutting-edge data analysis tools and techniques essential for identifying and solving complex business problems. This knowledge enhances your ability to improve performance, lower operation costs, increase project ROI, and make effective decisions. Furthermore, understanding generative AI fosters better team collaboration and risk management, crucial skills for any business analyst aiming to drive success and growth in an organization.
A6: The skill of “Can build a generative AI use case for own business/ company usage” will be very rare and hot in the market in the near future. If they know the business insight and can develop an AI tools into business use, will be very valuable “Innovate within the department” will come soon BA that don’t know coding will start to lose their value.
A7: After completing a course in generative AI, students typically come away with tangible projects or applications they've developed, such as a custom-designed chatbot or a unique AI-driven project. These creations showcase their practical skills in utilizing AI for real-world applications, demonstrating their ability to design, implement, and manage AI systems. Such projects serve as a portfolio piece, evidencing their proficiency in AI technologies and their application in business analysis, which can be valuable for career advancement or academic pursuits.
A8: The ideal candidates for this course are those interested in understanding how AI can be integrated into business practices. This course is suitable for individuals from any industry who wish to gain insights into AI applications in business contexts.
