We are pleased to announce our five-day Gen AI In-depth Course in conjunction together with Google! It was live between March 31 and 4 April 2025. It is it is now accessible in a self-paced guide to learning to those who are interested in knowing more about the basic technologies and methods that are behind Generative AI.
What's covered:

• Day One: Basic Models and Prompt Engineer: Explore the growth of Large Language Models (LLMs) starting with transformers and moving to advanced techniques such as fine-tuning, acceleration of inference as well as reasoning models. Learn the technique of prompt engineering to ensure the best LLM interaction.
• Day 2 Embeddings as well as Vector stores/databases. Learn about the theoretical foundations of embeddings as well as vector databases with embedding strategies and vector search algorithms as well as real-world applications using LLMs and the tradeoffs they bring.
• Day 3 Generational AI Agents Learn how to construct advanced AI agents through knowing their fundamental components and the development process that is iterative as well as the most recent developments in the area of agentspace.
• Day 4Domain-Specific LLMs Explore the development and use of specific LLMs such as SecLM and Med-PaLM. information from researchers who created them.
• Day 5MLOps are designed for Generative AI Learn to modify MLOps methods to Generative AI and leverage Vertex AI's foundational models as well as the generative AI applications.

Best of Luck!
This site is brought to you by Anant Nawalgaria Antonio Gulli, Mark McDonald, Polong Lin, Paige Bailey as well as many more participants through Google.

Day 1 (Foundational Large Language Models & Prompt Engineering)

Welcome to Day 1.

In this session, you'll discover the evolution of LLMs that range from transformers to advanced techniques such as fine-tuning and acceleration of inference. Also, you'll learn the process of prompt engineering, which is essential for the best LLM interaction as well as the evaluation of LLMs.

The first one will take users through how to get started using the Gemini 2.0 API. We will cover various methods for prompts, including how various factors affect prompts. In the next codelab you'll also be taught how to assess the performance to LLMs with the help of autoraters as well as well-structured output.

Day 1 Assignments:

1. Complete the Intro Unit - "Foundational Large Language Models & Text Generation":

Listen to the summary podcast episode to this unit.
To complement the podcast, read the "Foundational Large Language Models & Text Generation" whitepaper .

2. Complete Unit 1 - "Prompt Engineering":

Listen to the summary podcast episode This unit is for sale.
To complement the podcast, read the "Prompt Engineering" whitepaper .
Complete these codelabs on relute:
Prompting fundamentals
Structured and evaluated data
Make sure you phone verify the account before you start your account, this is required for account to be set up.
Want to have an Interactive dialogue ? Add the whitepapers NotebookLM
[Optional] Study an example find out how a renowned bank leveraged the latest technology and prompt engineering, as well as other content that were discussed in the assignments of day one to streamline the financial advisory process to achieve significant productivity improvements.

[Optional] Watch the YouTube livestream recordings . Paige Bailey will be joined by experts from Google and Google - Warren Barkley, Logan Kilpatrick, Kieran Milan, Anant Nawalgaria Irina Sigler, and Mat Velloso to discuss today's codes and readings.

Day 2 (Embeddings and Vector Stores/Databases)

Welcome to Day 2.

This morning, you'll learn about the fundamental underlying concepts behind embeddeddings and vector databases, and how they could be utilized to integrate the latest or specialist data in an LLM application. We will also look at their geometrical capabilities for categorizing and comparing textual information in addition to how to assess the embeddings.

Day 2 Assignments:

Complete Unit 2: "Embeddings and Vector Stores/Databases":

Listen to the summary podcast episode This unit is for sale.
To complement the podcast, read the "Embeddings and Vector Stores/Databases" whitepaper ..
Complete these code labs on relute:
Build a RAG question-answering system over custom documents
Explore Similarity of text with embedded embeddings
Build A neural classification network using Keras embedded embeddings.
Want to have an Interactive dialogue ? Add the whitepapers NotebookLM

(Optional) View this YouTube livestream video. Paige Bailey will be joined by highly experienced speakers from Google as well as Google employees Andre Araujo, Patricia Florissi, Alan Li, Anant Nawalgaria Xiaoqi Ren, Chuck Sugnet as well as Howard Zhou to discuss embeddings and data stores/vectors.

Day 3 (Generative Agents)

Welcome to Day 3.

In this session, you'll be able to construct complex AI agents through understanding their basic components and method of development that iteratively. Additionally, you'll learn about advanced architectures for agents and techniques, including multi-agent systems agents evaluation, and much more.

The Codelabs explain the process of connecting LLMs to systems already in place and with the world of. Discover the concept of function calling through providing SQL tools to chatbots (including an example based on Gemini 2.0's Live API ) to learn you can build a LangGraph system which can take orders at a café.

Day 3 Assignments:

Complete Unit 3a - "Generative AI Agents":

Listen to the summary podcast episode This unit is for sale.
To complement the podcast, read the "Generative AI Agents" whitepaper .
Complete these code labs on relute:
Talk Access to a database using the function of to a database with function.
Build an agentic or asynchronous ordering system inside LangGraph.
[Optional] Advanced 3b - "Agents Companion":

Listen to the summary podcast episode to this unit.
Read the advanced "Agents Companion" whitepaper .
Want to have an Interactive discussion ? You can try adding whitepapers NotebookLM
[Optional] Take a look at an example in which they describe how an industry-leading technology regulator utilized the agentic-generative AI technology to streamline the creation of tickets-to-code within software development. The result was an 2.5x increase in productivity.

[Optional] Watch the YouTube livestream recordings . Paige Bailey will be joined by the most knowledgeable speakers from Google and Google - Alan Blount, Antonio Gulli, Steven Johnson, Jaclyn Konzelmann Patrick Marlow, Anant Nawalgaria and Julia Wiesinger to discuss generative AI agents.

Day 4 (Domain-Specific LLMs)

Welcome to Day 4.

In this article, you'll explore the development and use of special LLMs such as SecLM and MedLM/MedPaLM. You'll get knowledge from the scientists who developed them.

In the Codelabs, you'll discover how to incorporate actual data into a model by using Google Search and then visualize the model using plotting tools through the Live API.You will also be taught how to tweak a customized Gemini model based on labels you create for solving custom-made tasks.

Day 4 Assignments:

Complete Unit 4 - "Domain-Specific LLMs":

Listen to the summary podcast episode to this unit.
To complement the podcast, read the "Solving Domain-Specific Problems Using LLMs" whitepaper .
Complete these codelabs on relute:

Tune a Gemini model for a custom task
Use Google Search data in generation.
Want to have an Interactive discussion ? Add the whitepapers NotebookLM .
[Optional] Watch the YouTube livestream recordings . Paige Bailey will be joined by knowledgeable speakers from Google Donny Cheung Scott Coull, Ewa Dominowska, Chris Grier, Anant Nawalgaria, and Karthik Raman who will discuss the specific domain models.

Day 5 (MLOps for Generative AI)

Welcome to Day 5.

Learn how you can adapt MLOps methods to Generative AI and make use of Vertex AI's software for building base models as well as Generative AI applications, such as AgentOps for applications that use agents.

Day 5 Assignments:

Complete Unit 5 - "MLOps for Generative AI":

Listen to the summary podcast episode This unit is for sale.
To complement the podcast, read the "MLOps for Generative AI" whitepaper ..
No codelab for today! During the livestream tomorrow, we will do a code walkthrough and live demo of goo.gle/agent-starter-pack This resource was created to assist in creating MLOps to Gen AI easier and accelerating the production process. You are encouraged to browse the repository prior to.
Want to have an Interactive dialogue ? Add the whitepapers NotebookLM

[Optional] Watch the YouTube livestream recordings . Paige Bailey will be joined by experts from Google and Google X - Sokratis Kartakis Gabriela Hernandez Larios Ivan Nardini, Anant Nawalgaria, Elia Secchi, Michael Styer and Saurabh Tiwary, to talk about the MLOps practice in Generative AI.

Bonus Assignment

Look over this bonus project It's not over! The notebook provides you with some additional things that you can accomplish using Gemini API. Gemini API that weren't covered throughout the course. The material isn't a complement to the podcast or whitepaper however it covers a few additional options you could find useful for creating Gemini API powered apps.