Back to Courses
Intermediate

AI Agentic Engineering Fundamentals

Learn to design and build AI-powered applications, agents, and multi-agent systems using Java. You'll implement RAG pipelines, function calling, agentic workflows, and production-ready patterns using Spring AI, LangChain4j, and Embabel.

10 Modules
58 Topics
Java Fundamentals required
Spring AI · LangChain4j · Embabel

Course Curriculum

  • How large language models work
  • Tokens, context windows & temperature
  • Prompt engineering fundamentals
  • Model capabilities & limitations
  • Choosing the right model for the task

  • Spring AI setup & model configuration
  • ChatClient & prompt templates
  • Streaming responses
  • Structured output with bean mapping
  • Multimodal inputs (image, audio)
  • Advisors & conversation memory

  • LangChain4j setup & chat models
  • AI Services & declarative interfaces
  • Structured prompts with @SystemMessage
  • Response parsing & type-safe output
  • Chat memory & message history

  • Defining tools & @Tool annotation
  • Tool execution & result injection
  • Parallel & sequential tool calls
  • Error handling in tool execution
  • Security considerations for tool use

  • Embeddings & semantic similarity
  • Vector stores — PGVector, Chroma, Weaviate
  • Document ingestion & chunking strategies
  • Retrieval pipeline & similarity search
  • Re-ranking retrieved results
  • Evaluating RAG quality

  • Conversation memory types — window, summary
  • Persistent memory with databases
  • Token budget management
  • Context compression techniques
  • Session isolation & multi-user memory

  • The agent loop — observe, plan, act
  • ReAct (Reasoning + Acting) pattern
  • Multi-step task decomposition
  • State management across steps
  • Human-in-the-loop checkpoints
  • Handling agent failures & retries

  • Embabel architecture & goals
  • Defining agents & their capabilities
  • Supervisor & worker agent pattern
  • Parallel agent execution
  • Agent communication & result aggregation

  • MCP protocol overview
  • Building an MCP server in Java
  • Exposing tools & resources via MCP
  • Consuming MCP servers from agents
  • MCP security & authentication

  • Observability — tracing LLM calls
  • Cost management & token budgets
  • Latency optimisation & prompt caching
  • Evaluating AI application quality
  • Testing AI applications with deterministic assertions