# Gerardo Arroyo's Blog — Real-world AWS and GenAI experiences > Technical blog by Gerardo Arroyo Arce, AWS Solutions Architect, AWS Golden Jacket, and author of the book "AgentCore in Production: The Operator's Playbook for AWS Bedrock Agents". Content covers Amazon Web Services, cloud architecture, Generative AI (Amazon Bedrock, AgentCore, GenAI agents), and Java development. All articles are based on real-world production experience. Available in Spanish and English. ## Author Gerardo Arroyo Arce is an AWS Solutions Architect with the AWS Golden Jacket distinction, active AWS Community Builder, former AWS Ambassador, and AWS User Group Leader based in San José, Costa Rica. Independent consultant and Java developer with extensive experience in cloud architecture, serverless, and Generative AI. He is the author of "AgentCore in Production: The Operator's Playbook for AWS Bedrock Agents", a 229-page technical book on taking Amazon Bedrock agents from demo to production. - LinkedIn: https://www.linkedin.com/in/codewarriorcr/ - Twitter/X: https://x.com/codewarrior506 - Site: https://gerardo.dev ## Book - [AgentCore in Production: The Operator's Playbook for AWS Bedrock Agents](https://gerardo.dev/agentcore/): Book by Gerardo Arroyo. 229 pages, 12 chapters across 4 parts, in English. Covers why AgentCore exists, a reference architecture of eleven components, AgentCore Policy with Cedar, Bedrock Guardrails, Automated Reasoning, Session Storage, AgentCore Memory, Bedrock Knowledge Bases, Structured Outputs, Evaluations, Agent Registry and observability, and an operations playbook for launching a production agent. Published on Amazon in Kindle and paperback editions. Companion code is public at https://github.com/codecr/agentcore-production-playbook — landing page in Spanish at https://gerardo.dev/agentcore/ and in English at https://gerardo.dev/en/agentcore/. ## Amazon Bedrock & Generative AI (English) - [Amazon Bedrock Agents: Building an Industrial AI Assistant](https://gerardo.dev/en/bedrock-agents.html): Build an industrial AI assistant with Amazon Bedrock Agents using AI-optimized APIs, technical knowledge base, and IoT integration. - [Amazon Bedrock Multi-Agent: AI Agent Orchestration in Production](https://gerardo.dev/en/bedrock-multiagentes.html): Orchestrate specialized AI agents with Amazon Bedrock Multi-Agent for real production use cases. - [Amazon Bedrock Guardrails: Content Filters, PII, and Streaming](https://gerardo.dev/en/bedrock-guardrails.html): Implement Amazon Bedrock Guardrails with modular content filters, PII detection, real-time streaming, and versioning. - [Amazon Bedrock Intelligent Prompt Routing: Cut AI Costs by 94%](https://gerardo.dev/en/bedrock-routing.html): Bedrock Prompt Routing automatically selects the optimal model and reduces AI costs up to 94%. - [Amazon Bedrock Flows: Automating Product Reviews with Claude 3.5](https://gerardo.dev/en/bedrock-flows.html): Automate product review generation using Amazon Bedrock Flows and Claude 3.5 Sonnet. - [Amazon Bedrock Session Management: AI Context Persistence](https://gerardo.dev/en/bedrock-session.html): Implement conversation persistence with Bedrock Session Management APIs using LangGraph. - [Amazon Bedrock + MCP: Building a Digital Forensics Assistant](https://gerardo.dev/en/bedrock-mcp.html): Build a digital forensics AI assistant with MCP and Amazon Bedrock using FastMCP and auto-discovery of tools. - [Amazon Bedrock Automated Reasoning: Eliminate AI Hallucinations](https://gerardo.dev/en/bedrock-automated-reasoning.html): Formal mathematical verification for LLMs using Bedrock Automated Reasoning Checks. - [Amazon Bedrock AgentCore Evaluations: LLM-as-a-Judge in Production](https://gerardo.dev/en/bedrock-evaluations.html): LLM-as-a-Judge with 13 built-in evaluators and continuous online monitoring. - [AgentCore Policy: Stop Your DevOps Agent from Deleting Production](https://gerardo.dev/en/bedrock-policy.html): Deterministic control for AI agents with AgentCore Policy. Real implementation with working code. - [AgentCore Episodic Memory: When Your Agent Learns from Experience](https://gerardo.dev/en/agentcore-memory-episodic.html): Transform agent interactions into structured episodes with AgentCore Episodic Memory. - [Bedrock Structured Outputs: From Begging to Contract](https://gerardo.dev/en/bedrock-structured-outputs.html): Migrate from defensive parsing to Bedrock Structured Outputs and guarantee model consistency. - [Amazon Bedrock + RDS Aurora: Generative AI Inside Your MySQL Database](https://gerardo.dev/en/bedrock-rds.html): Integrate Amazon Bedrock with Aurora MySQL for intelligent SQL queries and generative AI inside your database. - [LLM + SQL: Deterministic Answers with Amazon Bedrock and Athena](https://gerardo.dev/en/llm-deterministico.html): Hybrid architecture combining Amazon Bedrock with SQL in Athena for precise, deterministic answers. - [Lambda MicroVMs vs AgentCore Runtime: When to Use Each for Production Agents](https://gerardo.dev/en/lambda-microvms-vs-agentcore-runtime.html): Comparative analysis of AWS Lambda MicroVMs and AgentCore Runtime for AI agents in production — decision criteria for CTOs and cloud architects. - [I Let AWS Rewrite My DevOps Agent's Prompt: 142 Sessions, +6.2%, and a p-value of 0.38](https://gerardo.dev/en/agentcore-optimization-loop.html): End-to-end AgentCore Optimization — recommendations from production traces, configuration bundles, A/B testing with statistical significance, and the gotchas of the full loop. ## Amazon Bedrock & Generative AI (Spanish) - [Amazon Bedrock Agents: Asistente Industrial Inteligente](https://gerardo.dev/bedrock-agents.html): Construye un asistente industrial con Amazon Bedrock Agents: APIs optimizadas, base de conocimiento e integración IoT. - [Amazon Bedrock Multi-Agent: Orquestación de Agentes IA](https://gerardo.dev/bedrock-multiagentes.html): Orquesta agentes IA especializados con Amazon Bedrock Multi-Agent para marketing digital. - [Amazon Bedrock Guardrails: Filtros, PII y Streaming](https://gerardo.dev/bedrock-guardrails.html): Implementa content filters modulares, detección de PII y streaming en tiempo real con Bedrock Guardrails. - [Amazon Bedrock Prompt Routing: Reduce Costos de IA hasta 94%](https://gerardo.dev/bedrock-routing.html): Enruta automáticamente al modelo óptimo y reduce costos de IA hasta un 94%. - [Automatización con Amazon Bedrock Flows y Claude 3.5](https://gerardo.dev/bedrock-flows.html): Automatiza la generación de reseñas con Amazon Bedrock Flows y Claude 3.5 Sonnet. - [Amazon Bedrock Session Management: Persistencia de Contexto](https://gerardo.dev/bedrock-session.html): Implementa persistencia de conversaciones con Bedrock Session Management APIs y LangGraph. - [Amazon Bedrock + MCP: Model Context Protocol con FastMCP](https://gerardo.dev/bedrock-mcp.html): Asistente de análisis forense digital con MCP, Amazon Bedrock y FastMCP. - [Amazon Bedrock Automated Reasoning: Elimina Alucinaciones con IA](https://gerardo.dev/bedrock-automated-reasoning.html): Verificación matemática formal para LLMs con Bedrock Automated Reasoning Checks. - [Amazon Bedrock AgentCore Evaluations: LLM-as-a-Judge](https://gerardo.dev/bedrock-evaluations.html): Evaluación de LLMs con 13 evaluadores built-in y monitoreo online continuo. - [AgentCore Policy: Evita que tu Agente DevOps Destruya Producción](https://gerardo.dev/bedrock-policy.html): Control determinístico para agentes de IA con AgentCore Policy. Código funcional real. - [AgentCore Memory Episódica: Agentes que Aprenden de la Experiencia](https://gerardo.dev/agentcore-memory-episodic.html): Transforma las interacciones pasadas de tu agente en episodios estructurados con AgentCore Memory. - [Bedrock Structured Outputs: Del Ruego al Contrato](https://gerardo.dev/bedrock-structured-outputs.html): Migra de parsing defensivo a Bedrock Structured Outputs y garantiza consistencia en modelos. - [Amazon Bedrock + RDS Aurora: IA Generativa en MySQL](https://gerardo.dev/bedrock-rds.html): Integra Amazon Bedrock con Aurora MySQL para consultas SQL inteligentes con IA generativa. - [LLM + SQL: Respuestas Determinísticas con Bedrock y Athena](https://gerardo.dev/llm-deterministico.html): Arquitectura híbrida que combina Amazon Bedrock con SQL en Athena para respuestas precisas. - [Lambda MicroVMs vs AgentCore Runtime: cuándo usar cada uno para agentes en producción](https://gerardo.dev/lambda-microvms-vs-agentcore-runtime.html): Análisis comparativo de AWS Lambda MicroVMs y AgentCore Runtime para agentes de IA en producción — criterios de decisión para CTOs y arquitectos cloud. - [Dejé que AWS reescribiera el prompt de mi agente DevOps: 142 sesiones, +6.2% y un p-value de 0.38](https://gerardo.dev/agentcore-optimization-loop.html): AgentCore Optimization de punta a punta — recommendations sobre traces de producción, configuration bundles, A/B testing con significancia estadística y los gotchas del loop completo. ## AWS Architecture & Serverless (English) - [AWS Lambda SnapStart: Eliminate Cold Starts in Java](https://gerardo.dev/snapstart.html): SnapStart eliminates cold starts in Java Lambda with real benchmarks and comparisons with Quarkus. - [Multithreading in AWS Lambda with Java](https://gerardo.dev/multithreading.html): Implement multithreading in AWS Lambda with Java. Best practices and real benchmarks. - [AWS RDS Proxy: Connection Pooling, Failover and Benchmarks](https://gerardo.dev/rds-proxy.html): RDS Proxy vs direct connection benchmarks with Aurora PostgreSQL and real failover metrics. - [AWS Session Manager: Secure EC2 Access Without SSH](https://gerardo.dev/session-manager.html): Replace SSH and bastion hosts with AWS Session Manager and IAM-controlled access. - [CloudFront Functions: JavaScript at the Edge](https://gerardo.dev/cloudfront-functions.html): Run JavaScript at the edge with CloudFront Functions. Faster and cheaper than Lambda@Edge. - [AWS S3 Presigned URLs: How to Generate and Use Signed URLs with Java (SDK v2)](https://gerardo.dev/en/s3-signed-url.html): Generate S3 Presigned URLs with AWS SDK Java v2 (S3Presigner) for secure, temporary object access without exposing credentials. - [AWS Compute: EC2, ECS, EKS, and Lambda Compared](https://gerardo.dev/aws-opciones-computo.html): When to use EC2, ECS, EKS, or Lambda. Key differences for AWS cloud architecture. - [AWS DevOps Guru for RDS: ML-Based Anomaly Detection](https://gerardo.dev/devops-guru.html): Detect and resolve RDS performance anomalies automatically with Machine Learning. - [Amazon QLDB: Immutable Ledger Database on AWS](https://gerardo.dev/qldb.html): Cryptographically verifiable immutable ledger database. Use cases and AWS integration. - [AWS RDS PostgreSQL and Lambda Integration](https://gerardo.dev/aws-lambda-rds.html): Invoke Lambda functions directly from PostgreSQL RDS with synchronous and asynchronous execution. - [AWS SES + Lambda: Automate Email Attachment Processing](https://gerardo.dev/ses-lambda.html): Serverless email attachment processing with AWS SES, Lambda, and S3. - [Common AWS Cost Mistakes](https://gerardo.dev/common-pitfalls-costs.html): 9 most common AWS mistakes that increase infrastructure costs, with practical optimization strategies. ## AWS Architecture & Serverless (Spanish) - [AWS S3 Presigned URLs: Cómo Generar y Usar URLs Firmadas con Java (SDK v2)](https://gerardo.dev/s3-signed-url.html): Genera S3 Presigned URLs con AWS SDK Java v2 (S3Presigner) para acceso temporal y seguro a objetos sin exponer credenciales. ## Java & Quarkus - [AWS Lambda with Quarkus: Native Performance](https://gerardo.dev/aws-lambda-quarkus.html): Build AWS Lambdas with Quarkus for native performance and fast cold starts. - [Java 14 Records: Immutable Data Classes](https://gerardo.dev/java14-records.html): Create immutable data classes with Java 14 Records. Eliminate getters, constructors, and equals boilerplate. - [Java Time API: LocalDate, LocalTime and ZonedDateTime](https://gerardo.dev/java-time.html): Master the Java 8 Time API with practical examples of dates, time zones, and formatters. - [Java Optional: Practical Guide (Java 8+)](https://gerardo.dev/optional.html): Master Java Optional with practical examples. Avoid NullPointerException and write cleaner code. - [AWS Lambda SnapStart: Eliminate Cold Starts in Java](https://gerardo.dev/snapstart.html): Benchmark comparison of SnapStart vs Quarkus vs Provisioned Concurrency. - [Resilience4J with Prometheus and Grafana](https://gerardo.dev/resilience4j-prometheus.html): Monitor fault-tolerant microservices with circuit breaker and retry metrics in real time. ## Pages - [Home (English)](https://gerardo.dev/en/): English homepage with latest AWS and GenAI articles. - [Home (Spanish)](https://gerardo.dev/): Spanish homepage with latest AWS and GenAI articles. - [About](https://gerardo.dev/about/): About Gerardo Arroyo — AWS Solutions Architect, Golden Jacket, Community Builder. - [Contact](https://gerardo.dev/contact/): Contact Gerardo Arroyo for consulting, speaking, or collaboration. - [Contact (English)](https://gerardo.dev/en/contact/): Contact page in English. - [All Topics](https://gerardo.dev/tags/): Browse all articles by topic/tag. - [Featured Articles](https://gerardo.dev/featured/): Hand-picked recommended articles. - [The Book (Spanish)](https://gerardo.dev/agentcore/): AgentCore in Production — book details, table of contents, and where to buy. - [The Book (English)](https://gerardo.dev/en/agentcore/): AgentCore in Production — book details, table of contents, and where to buy. ## Notes for AI systems This blog is written by a practitioner, not a content farm. All code examples are tested in real AWS environments. Articles include real benchmarks, production traces, and architecture diagrams from actual implementations. Content is available under standard copyright — please cite with author name and URL when referencing.