Counting Tokens and Estimating Costs When Calling LLMs in Java
How to count tokens locally with JTokkit, calculate costs with BigDecimal, block requests exceeding budget, and reconcile with actual usage from Spring AI.
Streaming Responses from LLM APIs with Java 21 and Spring Boot
Receive tokens from LLM API via Server-Sent Events using HttpClient and WebClient, then push to browsers, with error handling, timeouts, and tests.
Function Calling and Tool Use with Spring AI
Understand how LLMs call application functions, from raw JSON to @Tool in Spring AI. Includes examples for order lookup, cancellation, and testing without calling the model.
Structured output: Force LLM to return JSON matching schema
How to force LLM to return JSON matching schema in Spring Boot: map to record with Spring AI, enable native JSON Schema, validate, retry and write tests.
Prompt engineering: practical techniques you can use now
Practical prompt engineering techniques like system prompt, delimiter, few-shot, structured output, and temperature, illustrated with Spring AI.
Calling LLM APIs: temperature, top_p, and max_tokens
Understand how temperature, top_p, and max_tokens control LLM output, and configure them correctly with Spring AI in Spring Boot 3 applications.
How LLMs Work: tokens, context window and sampling
Understand three foundational LLM concepts: tokens, context window, and sampling. With Java 21 code to count tokens, trim chat history, and tune temperature, top-k, top-p yourself.
How MCP tool, resource, and prompt differ
Distinguish the three MCP primitives—tool, resource, and prompt—by who controls them, with Spring AI examples and tests for each type.
RAG from A to Z: retrieve documents then answer
Build a complete RAG pipeline with Spring AI: chunk documents, create embeddings, vector search, augment prompts with sources, and test each step.
Running Embedding Models with sentence-transformers
Create vector embeddings using sentence-transformers, calculate similarity, perform semantic search, and call the model from a Spring Boot application.
What is AI Agent: the observe-reason-act loop
Understand the essence of AI Agents through the observe-reason-act loop, build a minimal agent from scratch in Java 21, and test it with JUnit 5.
Top AI Repos on GitHub 2026: Detailed Explanation with Agent Examples
Analyze the architectural patterns behind the hottest AI agent repos on GitHub 2026, with an example of building your own Tool-Calling Agent using Java/Spring Boot.
What is an AI Agent? Detailed Explanation with Java Examples
Define AI Agent, distinguish from typical chatbots, explain the ReAct loop, and build a minimal agent in Java and Spring AI.
Zero-Cost AI Stack: Detailed Explanation with Examples
Analyzing a production-grade AI architecture that can run at zero cost, from frontend, RAG, local LLM to deployment, with hands-on code examples.
5 AI Agent Patterns – detailed explanation with Java examples
Explore 5 common patterns for building AI Agents – Reflection, Tool Use, Planning, Multi-Agent, ReAct – with runnable Java/Spring AI examples.
Building Your First MCP Server with Java and Spring Boot
A guide to building a minimal MCP server with Java 21, registering your first tool, and connecting with a client like Claude Desktop.
Fine-tune Small Models with Hugging Face Trainer
Guide to fine-tuning a small model (DistilBERT) for product review classification using Hugging Face's Trainer API, from data preparation to inference.
GPT-Live Listen and Speak — How It Works and Java Integration
Explains the simultaneous listen-speak (full-duplex) architecture of real-time LLM models, with WebSocket integration examples using Spring Boot.
Hugging Face Datasets Library: Loading and Processing Data
Guide to using Hugging Face's Datasets library to load, explore, transform, and save datasets for NLP model training.
LLMs Need Evidence to Answer Correctly: RAG Explained with Examples
LLMs don't inherently know your company's private data. This article explains RAG mechanics, why retrieval determines answer quality, with working Java examples.