José Adrián Alemán Rojas

Systems architecture, backend platforms, and technical leadership.

I'm a software engineer, technical lead, and architect with 25+ years building Java and Spring Boot backends, microservices, and cloud-native systems for enterprise and regulated domains. As founder of ITJobOpportunities, I ship the hiring platform, recruiter ATS, and Code Training Lab for technical screening across Costa Rica and LATAM. In parallel I build applied AI systems in production: RAG and product chat, LLM orchestration, software agents, and engineering skills for AI coding agents. I publish the architecture notes and sketchnotes behind that work at my-tech-profile.dev. I care most about dependable delivery, clear API and distributed-systems design, and teams that can extend what we ship with confidence.

Insights

• A Firewall for AI Agents: Enforce Authority at Every Tool Call
• MCP talks. OWASP names the damage. ACS is the brake
• MCP is a Host, a Client, and a Server. Stop Mixing Them Up
• From ETL, ELT, and EtLT to Agents: What's Actually Changing in Data Engineering
• How I would build Agentic RAG in LangChain4j
• Agentic RAG is a graph with a grade, not a fancier prompt
• Loop Engineering isn't Dead
• Who disappears first with AI — frontend or backend?
• What comes after microservices is an agent with a budget
• Spring AI's AutoMemoryTools: Letting the Agent Decide What's Worth Keeping
• What 99 insights on my-tech-profile.dev actually add up to
• Build a production-ready AI PR reviewer in Java with LangChain4j
• SuperPowers: A Skill Framework
• Vibe Coding — Vulnerability as a Service
• What Is Actually Cheaper for Orchestrating Coding Agents: APIs, Codex CLI, or Cursor CLI?
• DB-Scheduler: A Persistent, Cluster-Friendly Scheduler for Java
• API Mocking and Testing With Microcks
• A Practical Guide to Java Virtual Threads With JMS Listeners on Java 25 and Spring Boot 4
• Is n8n better than Spring AI or LangChain4j?
• Securing MCP servers in agent workflows
• Prompt Engineering Techniques with Spring AI: What Actually Works in Production Java
• A Goal the Agent Cannot Fake
• The Evolution Of Modern Software Architectures
• Spring AI 2.0 vs LangChain4j 1.18: same research agents, different orchestration
• Google Gemini with Spring AI 2.0
• Agent2Agent on Spring Boot: one specialist, one orchestrator, no megaprompt
• Use of LLMs to Learn Complex Topics
• Building Resilient Data Pipelines: Practical Patterns That Survive Production
• System Prompt Leakage vs. Prompt Injection: Two Different Attacks, Two Different Defenses
• Create a AI Video from Codebase
• SRE Best Practices For Production Alerting
• LLM Integration with Apache Camel
• This One Spring Data JPA Pattern Cleaned Up to 3 Years of Repository Debt
• Microservices Architecture: 7 Engineering Decisions That Determine Success or Failure
• Stop Giving Agents Your Feign Clients: Govern Tools Behind MCP
• Application Startup Tracking In Spring Boot
• What Spring AI 2.0 Actually Gives You for Transactional Agent Workflows
• Don't write prompts, run loops
• Thirty years of Java, and I've been there for most of it
• Spring Boot 4.0 changed how HttpStatus prints
• Deep Dive on Dokimos for AI Spec Testing in Java
• Spec-Driven Development (Part 2)
• ACK as a AI Start Response
• Anti-Patters of Microservices Architecture
• My honest take on the "1 million notifications" system design question
• AGENTS.md Makes Your Java Codebase AI-Agent Ready AGENTS.md Makes Your Java Codebase AI-Agent Ready
• DuckDB as an AI Data Layer
• AI Integration to AI Engineering
• Mastering Java Loops: Under the Hood, Memory Pitfalls, and High-Scale Optimizations
• Stop Making Your AI Agent Re-Read Your Entire Codebase
• Database Normalization, ACID, and Slowly Changing Dimensions
• Spring Boot 4 vs Spring Boot 3
• Ontology Knowledge Fusion: Why I Stopped Trusting Pure RAG for Job Fit
• Getting Started With RabbitMQ in Spring Boot
• Centralizing Chaos Monkey at Scale — What I Learned Building a Command Relay
• Picking Between LangChain4j 1.15 and Spring AI 2.0
• Why AI-Generated Code Is Making Regression Testing More Important
• Deep dive on Spring Modulith
• The New Senior Developer Job Description
• Multi-Agent Orchestrator and Workers in Spring AI 2.0
• AI-Native Engineering Culture
• Deep Dive on Spring AI 2.0
• RAG to Agentic RAG to Agent Memory
• Bean Background Initialization in Spring Framework
• When to use Rest API, WebSockets and grpc
• Stop Copy-Pasting Agent Rules Into Every Repo
• AI Broke Your Definition of Done
• Chaos Engineering with Chaos Monkey
• Self-Correcting Structured Output in Spring AI 2.0
• GenAI Isn't Solving the Problem Most Development Teams Actually Have
• Deep Dive in DokDeploy
• Measure your RAG with RAGAS
• Git Worktrees
• Audit Trails Make Systems Easier to Trust
• LLM Architecture Explained
• Claude Fable 5 Changes the Model-Selection Question
• Orchestrating Subagents in Spring AI: What I Learned Building It the Hard Way
• AI Spec-Driven Development
• What Actually Is an LLM?
• Building Code Training Lab: a self-hosted multi-language coding platform
• AL & ML Concepts Explained
• MergeWatch + GitHub Actions: A Setup That Actually Made Our PR Reviews Useful
• AI-Generated Code Testing Strategies
• The Algorithmic Yes-Man But Why?
• Dynamic Workflows With Opus 4.8
• When Your AI Agent Fails Mid-Task: Temporal vs LangGraph vs Axon
• The Different Kinds of RAG, and How I've Been Using LangChain4j to Build Them
• Deep Dive on AI Harness Engineering
• MartinLoop for AI Workflows
• SurfSense vs Onyx: Two Very Different Bets on Personal AI
• LlamaIndex RAG Pipelines
• Building a Verification Layer That Catches Hallucinations
• Learn Agentic AI: Here's Where I'd Actually Start
• AI Governance and Safety: What I've Learned Building Systems That Actually Have to Be Accountable
• Introduction to Github Actions
• Filtering the Spam Before It Gets Into Your Vector Store
• I Built a Small POC to Compare FreeBuff and CodeBuff
• From APIs to Actions: Rethinking Back-End Design for Agents
• The Hidden Bottlenecks That Break Microservices in Production
• How I Got Our API Response Times Under 100ms With Redis and Spring Boot
• Skills vs. Prompts
• Building with OpenWA Java Whatsapp Bots
• Context-Aware Authorization for AI Agents
• Using Anthropic's Agent Skills in Spring AI
• Caddy to the Rescue: Why I Stopped Fighting Nginx and Never Looked Back
• Building a profesional branding for you
• Software Design Pattens in Java
• The Art of Token Frugality in Generative AI Applications
• Spring Boot Observability with Open Telemetry and DynaTrace
• Observability in Apache Camel Applications

Sketchnotes

• GitHub Actions
• Spring Boot 3 to 4 Migration
• Classic RAG vs Agentic RAG
• LangGraph4j vs LangChain4j
• Architect Secure AI Agents
• RAG vs Context-Augmented Generation
• Python vs Java AI Tools
• Proof of Humanity
• Cursor Origin vs GitHub
• How to Think Like an AI Architect
• llmfit · right-size the local LLM
• Claude · Codex from your phone
• n8n vs LangChain4j 1.18
• Java Records · Use Them Well
• When to Vibe · When to Spec
• LangChain4j 1.18 — Builders · Domain · Observability
• Agentic Era Languages · 2026
• MCP vs Agent2Agent (A2A)
• Prompt Engineering vs Loop Engineering
• Agent Memory + MCP
• What is a Forward Deployed Engineer
• Apache Camel vs Spring Integration
• MCP Client & Server · Spring AI 2.0
• Static Analysis in Spring Boot
• RAG vs Fine-Tune
• Claude Opus 5
• Avoid Agent Hallucinations · Spring AI 2.0
• AI Hype Versus Reality
• Kimi K3 vs DeepSeek V4 Pro
• RAG with LangChain4j & Spring AI 2.0
• SQS · RabbitMQ · Kafka
• How to Become a Java AI Engineer
• AI Guardrails
• Integrate with MCP
• When to Use an Agent
• Popular AI Models in 2026
• How an AI Spec Should Look
• Java 26 Roadmap
• Microservices vs Miniservices
• AI Native Harness
• GPT-5.6
• When to Vibe · When Not To
• Grok 4.5
• Voice Agent Architecture — Hear, Think, Speak
• How LLMs Generate Text — Inference Pipeline
• LangChain4j 1.15 vs Spring AI 2.0
• RAG Pipeline vs Agentic RAG
• Engineering Profile Sketchnote
• Spring Boot 4.1