Tutorials and deep-dive implementations covering artificial intelligence and machine learning concepts in Java and other languages.
25 interview questions on RAG, embeddings, tool calling, MCP, agents, evaluation and cost control for Java developers, with short answers in accordions and runnable proof for the tricky ones.
Jev is a hosted TypeSafe AI API that answers yes/no, pick-one and score questions about your data. Here is the Spring AI starter, how to wire it, and a confidence-gated model router, verified offline.
Package an agent's know-how as SKILL.md folders and let Spring AI load them on demand with one tool. What the model sees, what a load costs, and the edge cases I reproduced.
Spring AI 2.1.0-M1 adds VectorStore.upsert() that takes vectors you already have. Here is what it does, which stores support it, and how to move between stores without calling the embedding model again.
Spring AI 2.1.0-M1 can talk to OpenAI's Responses API. One property switches the model bean, and new options add hosted web search, code interpreter and remote MCP. What I verified, what I could not measure.
Spring AI 2.1.0-M1 models an assistant turn as an ordered list of parts. Here is what the model looks like, how streaming marks partial parts, and where reasoning is silently lost today.
Build a semantic cache for LLM calls in Spring Boot with Spring AI and Redis vector search, then measure threshold, hit rate, wrong answers and savings on 600 replayed requests.
Run a DistilBERT sentiment model from Java with ONNX Runtime and DJL: tokenizing, identical answers by two routes, batching, a Spring Boot endpoint, and measured latency.
Jlama runs a 4-bit TinyLlama inside the JVM on the JDK 25 Vector API. Measured tokens per second, streaming, the token-limit surprise, temperature, and a same-machine comparison with Ollama.
Two Java agents talk over the A2A protocol: the agent card, a raw SendMessage call, blocking versus streaming, input-required follow-ups and an agent calling another agent, with every message shown and a table comparing A2A with MCP.