Enterprise AI on the JVM
Despite what you may have heard, the JVM is where you want to build if you are serious about Enterprise AI. We will teach you to embrace software engineering discipline to build safe AI in Kotlin and Java that generates value for your organization.
Description
Every organization wants to make AI work, but FOMO is not a strategy. Study after study shows even big companies are having trouble making AI work for them. There are a lot of reasons. Leaders have no clear business objectives. They lack a data strategy. They also fail to understand that context and structure are essential to the success of enterprise AI at scale.
While Python is great for computation and experimentation and therefore a great fit for machine learning, it's not great for context and structure. This is where JVM languages like Java, which has dominated the enterprise for decades, and more recently Kotlin excel. Enterprise AI on the JVM teaches you how to build robust, production-grade AI at scale on a proven, reliable platform.
What makes this course different
This is one of the first courses of its kind because Python had a first mover advantage in the transition from machine learning to AI. It's only recently that businesses are starting to discover the limits of Python at enterprise scale that have hindered it for decades while Java, C#, and more recently TypeScript dominate. Enterprises have literally invested billions of dollars in JVM applications, and the JVM has risen to the occasion to offer some amazing options for AI applications in the enterprise that can leverage those investments.
Enterprise AI on the JVM begins with an introduction to AI concepts, the software engineering principles we have honed for decades that are essential to production AI, and the strengths the JVM offers over Python for the enterprise. You will then learn four frameworks that make it easy to bring all these ideas together:
- Spring AI
- LangChain4j
- Embabel
- Koog
When you finish this course, you will be an expert at building AI systems in production at scale. You will even be able to apply your skills across technology stacks. You will build better architected AI systems in TypeScript, C#, and of course Python if you have to.
Course Syllabus
Lesson 1: Principles of AI
- Large Language Models
- Tokens: The Currency of AI
- Prompts and Prompt Templates
- Reasoning, Inference, and Thinking
- Reinforcement Learning
- Context Engineering
- Embeddings
- Retrieval Augmented Generation (RAG)
- Agents
- Tools
- Model Context Protocol (MCP)
- Guardrails
- Caching
- Evals
Lesson 2: AI in the Enterprise
- The Problem with Python
- The Dominance of Java and Power of Kotlin
- Type Safety
- Domain-Driven Design
- Loose Coupling
- Patterns in the Enterprise
- Security and Validation
- Observability
- Testability
- AI Only When Necessary
Lesson 3: Spring AI
- A Simple Chat Client
- Advisors
- Structured Output Converters
- Multimodality
- Memory
- RAG
- Tool Calls
- MCP
- Evals
- Guardrails
- Observability
- Developing with Docker Compose
- Testing
- Deployment
Lesson 4: Langchain4j
- A Simple Chat Client
- AI Services
- Multimodality
- Memory
- RAG
- Tool Calls
- MCP
- Evals
- Guardrails
- Observability
- Testing
- Deployment
Lesson 5: Building Agents with Embabel
- Where Embabel Excels
- Embabel Concepts
- Goal-Oriented Action Planing: Embabel's Killer Feature
- Your First Agent
- Agent Flows with Annotations
- Agent Flows with DSL
- Parallelism
- Using Tools and MCP
- Testing
- Observability
- Deployment
Lesson 6: Building Agents with Koog
- Where Koog Excels
- Koog Concepts
- Multiplatform Deployment: Koog's Killer Feature
- Your First Agent
- Strategy Graphs
- Parallelism
- Using Tools and MCP
- Testing
- Observability
- Deployment
Lesson 7: Agent Patterns
- Why Patterns Matter
- Applying Longstanding Patterns to AI
AI is changing the game and fast, and even the biggest companies in the world are having trouble making AI work for them. I want to help you harness AI in a serious way to be productive and build great things.