Artificial Intelligence (AI)

table of contents

---

• Artificial Intelligence (AI)
  • Notes for Oreilly Courses
  • References
  • Why AI for Java Developers?
    • Items Covered
  • Spring AI
  • Using OpenAI with spring-ai

---

Notes for Oreilly Courses

• Using AI as Personal Learning Coach
• vonage-learning
• GitHub Copilot
• Skills Plan: AI Agents
• Spring AI
  • Introduction
• Claude
• MCP
• OpenCode and GitAI
• Comparisions

References

• Text
  • Agent Skills Vs MCP Vs Prompts Vs Projects Vs Subagents :A Comparative Analysis
• Video
  • Harvard CS50’s Artificial Intelligence with Python – Full University Course
  • All Machine Learning algorithms explained in 17 min
  • How JP Morgan Built An AI Agent for Investment Research with LangGraph - LangChain Interrupt

Why AI for Java Developers?

• Industry-wide AI Adoption?
  • Retail
  • Healthcare
  • etc
• Enhance Problem Solving Abilities
• Stay Competitive in the Job Market

Items Covered

• Intro to LLM OpenAPI ChatGPT
• Chat based AI apps using OpenAI APIs
• Prompt Engineering
• Generate Structured Data with OpenAI
• Function Calling using Tools with OpenAI
• RAG Based Q&A
• Multimodality : Images, Vision and Audio with OpenAI
• Run LLMs locally and build AI apps interact with them.

Spring AI

• Module for building applications using LLM for easy AI integration into Spring Framework
• Builds Enterprise-Grade AI applications
• spring-ai docs here

Using OpenAI with spring-ai

• Account Setup
  • Create OpenAI account
  • create APIkey and store it as env variable
  • check the playground and do compare by asking same question - What is an LLM?
    • 4o vs 4o-mini
    • 4o gives options
    • 4o-mini assumed we asked for Large Language Model
• Install Java 24 using sdkman sdk list java and sdk install java 24.0.1-tem
• clone instructor’s project here
• org.springframework.ai:spring-ai-starter-model-openai and property spring.ai.openai.api-key=${OPENAI_KEY}
• ChatClient calls internally OpenAI chat completion endpoint here
• Promt, Tokens and Tokenizations
  • user prompt -> tokenizations -> Processing using Transformer Architecture -> Generate Tokens -> De-tokenization -> output
• importatnt params - temprature (lower is more relvant response), max_completion_tokens
• Prompt - User Message + System Message + Assistant Message
  • system message like *SystemMessage(“You are a helpful assistant, who can answer java based questions. For any other questions, please respond with I don’t know in a funny way!”);
• Streaming ChatClient response (Async - Non blocking) using Flux in response instead of sync response Object
• conversation - What’s my name return don’t know an

ChatClient.builder().build()
.prompt()
.user(userInput.prompt())
.system(systemMessage)
.call()
.content()

• SpringAdvisors