Application Development with LLMs on Google Cloud
(1 day)
In this course, you explore tools and APIs available on Google Cloud for integrating large language models (LLMs) into your application. After exploring generative AI options on Google Cloud, next you explore LLMs and prompt design in Vertex AI Studio. Then you learn about LangChain, an open-source framework for developing applications powered by language models. After a discussion around more advanced prompt engineering techniques, you put it all together to build a multi-turn chat application by using LangChain and the Vertex AI PaLM API.
Course Objectives
- Explore the different options available for using generative AI on Google Cloud
- Use Vertex AI Studio to test prompts for large language models
- Develop LLM-powered applications using LangChain and LLM models on Vertex AI
- Apply prompt engineering techniques to improve the output from LLMs
- Build a multi-turn chat application using the PaLM API and LangChain.
Audience
Application developers and others who wish to leverage LLMs in applications.
Prerequisites
Completion of Introduction to Developer Efficiency on Google Cloud or equivalent knowledge.
Course Outline
Module 1: Introduction to Generative AI on Google Cloud
- Vertex AI on Google Cloud
- Generative AI options on Google Cloud
- Introduction to course use case
Module 2: Vertex AI Studio
- Introduction to Vertex AI Studio
- Available models and use cases
- Designing and testing prompts in the Google Cloud console
- Data governance in Vertex AI Studio
- Lab: Exploring Vertex AI Studio
Module 3: LangChain Fundamentals
- Introduction to LangChain
- LangChain concepts and components
- Integrating the Vertex AI PaLM APIs
- Question/Answering Chain using PaLM API
- Lab: Getting Started with LangChain + Vertex AI PaLM API
Module 4: Prompt Engineering
- Review of few-shot prompting
- Chain-of-thought prompting
- Retrieval augmented generation (RAG)
- ReAct
- Lab: Prompt Engineering Techniques
Module 5: Creating Custom Chat Applications with Vertex AI PaLM API
- LangChain for chatbots
- Memory for multi-turn chat
- Chat retrieval
- Lab: Implementing RAG Using LangChain