Spring AI 2 - Fundamentals

Learn the core Spring AI chat API, set up a Maven project with springweb and olama, and master four chat patterns from simple prompts to structured output with provider-agnostic coding.

Learn prompt engineering by defining a system persona and user messages, then invoke a chat client to explain topics such as microservices and Java.

Explore chat options across different LLM providers, configure max tokens and other controls in Visual Studio Code, and see how temperature, top P, and penalties shape responses.

Learn to build and render a prompt template in Visual Studio Code, pass topic values to an LLM, and reuse templates for multi-variable prompts with placeholders.

Build a structured LLM output converter in Java using a prompt template and format mapping to return an actor and their films as a Java object.

Explore streaming in Spring AI by swapping the blocking call with stream, delivering LLM tokens to the client via SSE emitter, response body emitter, and streaming response body.

Build streaming chat with prompt templates in Spring AI 2 Fundamentals, defining a two-parameter prompt (topic and style) and enabling multi-turn conversations via server-sent events.

Explore alternative emitter types in Spring AI 2 - Fundamentals, including SSE emitter, response body emitter, and streaming response body, and compare push-based versus pull-based streaming and their use cases.

See how tool calling lets LLMs invoke external APIs to obtain real-time data, turning chatbots into agents. Explore static and dynamic tool declarations that drive the request flow.

Master static tool calling in Spring AI 2 fundamentals, configuring grok API keys and OpenAI, compare tools method versus tool names, and prepare for dynamic tool calling next lesson.

Explore dynamic tool calling by defining an API, building tools such as a random number generator and a reverse string tool, and registering them to invoke the LLM.

explore conversation memory for AI chats, enabling context across interactions and coding demos while balancing token costs with explicit history, window memory, summary memory, semantic vector memory, and knowledge graphs.

Demonstrates building a chat history pattern with an in-memory H2 database, JPA repositories, and Lombok entities to store and retrieve conversation history for LLM prompts.

Explore session based and window based memory patterns for chat histories, using conversation IDs to isolate context and summary memory to compress old messages.

Discover tool context, a powerful pattern that blends memory, tools, and context to build a personalized assistant while keeping server-side data out of the LLM.

Learn to implement a Redis-backed semantic cache for chat memory, leveraging embeddings and a vector store to speed responses. Set up cloud Redis, OLAMA, and OpenAI options.

Learn the basics of RAG, retrieval augmented generation, grounding LLM responses by retrieving relevant documents from your knowledge base through vector embeddings and semantic search, then composing grounded answers.