https://brunovicco.github.io/technical-articles/en/Notes on building, observing and governing AI systems. From agent architectures to the decisions that need to hold up in production.enhttps://brunovicco.github.io/technical-articles/en/articles/quem-esta-chamando-a-tool/https://brunovicco.github.io/technical-articles/en/articles/quem-esta-chamando-a-tool/Tue, 27 Oct 2026 12:00:00 GMTIn the previous article on autonomy and authority, I explored the difference between allowing an agent to choose a task's next steps and granting it authority to produce external effects. I used the Controlled Autonomy Lab as an…https://brunovicco.github.io/technical-articles/en/articles/agentic-sdlc-como-plataforma-o-que-precisa-existir-entre-o-coding-agent-e-o-merge/https://brunovicco.github.io/technical-articles/en/articles/agentic-sdlc-como-plataforma-o-que-precisa-existir-entre-o-coding-agent-e-o-merge/Thu, 22 Oct 2026 12:00:00 GMTWhen a coding agent starts participating in development, it enters a process that already has identity, repositories, dependencies, execution environments, tests, branch policies, CI/CD, and rules for promoting changes.https://brunovicco.github.io/technical-articles/en/articles/autonomia-nao-e-autoridade-onde-termina-a-decisao-do-agente-e-comeca-o-direito-de-agir/https://brunovicco.github.io/technical-articles/en/articles/autonomia-nao-e-autoridade-onde-termina-a-decisao-do-agente-e-comeca-o-direito-de-agir/Mon, 19 Oct 2026 12:00:00 GMTIn the Controlled Autonomy Lab, I implemented the same incident-analysis task using six architectures with different degrees of control and autonomy. The experiment's most autonomous pattern receives five tools and decides which to…https://brunovicco.github.io/technical-articles/en/articles/dora-developer-experience-e-coortes-como-eu-mediria-o-impacto-da-ia-em-engenharia/https://brunovicco.github.io/technical-articles/en/articles/dora-developer-experience-e-coortes-como-eu-mediria-o-impacto-da-ia-em-engenharia/Mon, 12 Oct 2026 12:00:00 GMTPart 2 of 2 - Measuring engineering in the age of agentshttps://brunovicco.github.io/technical-articles/en/articles/tokens-como-metrica-de-produtividade-onde-essa-leitura-quebra/https://brunovicco.github.io/technical-articles/en/articles/tokens-como-metrica-de-produtividade-onde-essa-leitura-quebra/Fri, 09 Oct 2026 12:00:00 GMTAt a company where I worked, someone proposed using token consumption as one of the indicators for tracking LLMs' impact on developer productivity. The executive's reasoning started with data already available: the provider allowed…https://brunovicco.github.io/technical-articles/en/articles/langgraph-agno-ou-crewai-como-escolher-a-abstracao-que-combina-com-o-problema/https://brunovicco.github.io/technical-articles/en/articles/langgraph-agno-ou-crewai-como-escolher-a-abstracao-que-combina-com-o-problema/Tue, 29 Sep 2026 12:00:00 GMTAfter discussing continuity difficulties in hybrid architectures, it is worth examining a more basic question: when would I start with LangGraph, Agno, or CrewAI? The three frameworks overlap, and all have evolved considerably. A…https://brunovicco.github.io/technical-articles/en/articles/brasil-ia-e-infraestrutura-regulada-o-proximo-desafio-da-adocao-corporativa/https://brunovicco.github.io/technical-articles/en/articles/brasil-ia-e-infraestrutura-regulada-o-proximo-desafio-da-adocao-corporativa/Tue, 22 Sep 2026 12:00:00 GMTBrazil is beginning to occupy a position that is hard to ignore in the global artificial intelligence market. According to data released by OpenAI in August 2026, the country is among ChatGPT's three largest markets by weekly active…https://brunovicco.github.io/technical-articles/en/articles/langgraph-agno-na-mesma-aplicacao/https://brunovicco.github.io/technical-articles/en/articles/langgraph-agno-na-mesma-aplicacao/Tue, 22 Sep 2026 12:00:00 GMTAfter dealing with an architecture where LangGraph and Agno shared responsibilities without a sufficiently clear boundary, the decision in my case was ultimately to migrate the entire solution to LangGraph.https://brunovicco.github.io/technical-articles/en/articles/seu-agente-esqueceu-a-conversa-ou-sua-arquitetura-perdeu-a-identidade-da-execucao/https://brunovicco.github.io/technical-articles/en/articles/seu-agente-esqueceu-a-conversa-ou-sua-arquitetura-perdeu-a-identidade-da-execucao/Wed, 16 Sep 2026 12:00:00 GMTAfter examining state and continuity in the first part of this series, one of the most frequent causes of strange behavior deserves separate attention: identity. In a stateful system, a request continues the previous story only if the…https://brunovicco.github.io/technical-articles/en/articles/multiagentes-quebram-no-estado-antes-de-quebrarem-no-prompt/https://brunovicco.github.io/technical-articles/en/articles/multiagentes-quebram-no-estado-antes-de-quebrarem-no-prompt/Tue, 15 Sep 2026 12:00:00 GMTThis series grew out of two very concrete situations. In one professional experience, I inherited a POC for a multi-agent architecture in which a supervisor built with LangGraph delegated work to agents implemented with Agno. The…https://brunovicco.github.io/technical-articles/en/articles/6-padroes-de-arquitetura-llm-aprendizados-de-234-celulas-experimentais-sobre-controle-e-autonomia/https://brunovicco.github.io/technical-articles/en/articles/6-padroes-de-arquitetura-llm-aprendizados-de-234-celulas-experimentais-sobre-controle-e-autonomia/Tue, 01 Sep 2026 12:00:00 GMTThe idea for this study emerged while I was taking an Anthropic learning path, part of which covers choosing between augmented LLMs, workflows, and agents based on two aspects that seem simple but have important consequences in…https://brunovicco.github.io/technical-articles/en/articles/um-trace-varios-agentes-observabilidade-distribuida-entre-a2a-e-mcp/https://brunovicco.github.io/technical-articles/en/articles/um-trace-varios-agentes-observabilidade-distribuida-entre-a2a-e-mcp/Mon, 17 Aug 2026 12:00:00 GMTWhile developing applications composed of multiple agents, I encountered a recurring observability problem. A request would start in an orchestrator, move to another agent over A2A, pass through an MCP service, and return to the…https://brunovicco.github.io/technical-articles/en/articles/risco-de-concentracao-em-ia-generativa/https://brunovicco.github.io/technical-articles/en/articles/risco-de-concentracao-em-ia-generativa/Mon, 13 Jul 2026 12:00:00 GMTA governance analysis of resilience, redundancy, and the cost of diversifying AI providers.https://brunovicco.github.io/technical-articles/en/articles/harness-engineering-como-tornar-agentes-confiaveis/https://brunovicco.github.io/technical-articles/en/articles/harness-engineering-como-tornar-agentes-confiaveis/Mon, 06 Jul 2026 12:00:00 GMTAgent = Model + Harnesshttps://brunovicco.github.io/technical-articles/en/articles/spec-driven-development-com-ia-cinco-anti-padroes-e-o-fluxo-que-torna-a-resposta-rastreavel/https://brunovicco.github.io/technical-articles/en/articles/spec-driven-development-com-ia-cinco-anti-padroes-e-o-fluxo-que-torna-a-resposta-rastreavel/Thu, 02 Jul 2026 12:00:00 GMTPart 5 of 6 — Series: Generative AI in a Regulated Financial Institutionhttps://brunovicco.github.io/technical-articles/en/articles/arquitetura-onde-a-politica-vira-controle-tecnico-e-por-que-a-residencia-de-dados-define-o-desenho/https://brunovicco.github.io/technical-articles/en/articles/arquitetura-onde-a-politica-vira-controle-tecnico-e-por-que-a-residencia-de-dados-define-o-desenho/Tue, 30 Jun 2026 12:00:00 GMTFrom policy to technical control - Part 4 of "Generative AI in a Regulated Financial Institution".https://brunovicco.github.io/technical-articles/en/articles/a-conta-da-ia-chegou-limitar-o-uso-e-a-resposta/https://brunovicco.github.io/technical-articles/en/articles/a-conta-da-ia-chegou-limitar-o-uso-e-a-resposta/Tue, 23 Jun 2026 12:00:00 GMTRouting, throughput, and governancehttps://brunovicco.github.io/technical-articles/en/articles/frameworks-na-pratica-nist-ai-rmf-iso-42001-owasp-e-gaig-traduzidos-em-controle-e-evidencia/https://brunovicco.github.io/technical-articles/en/articles/frameworks-na-pratica-nist-ai-rmf-iso-42001-owasp-e-gaig-traduzidos-em-controle-e-evidencia/Tue, 23 Jun 2026 12:00:00 GMTThere is a crucial difference between citing a framework in a Board presentation (or merely putting it in a policy) and implementing it so that an external auditor can verify it. The first is easy and fits on a slide; the second…https://brunovicco.github.io/technical-articles/en/articles/o-labirinto-regulatorio-brasileiro-o-que-bacen-cvm-anbima-e-lgpd-exigem-antes-do-primeiro-prompt/https://brunovicco.github.io/technical-articles/en/articles/o-labirinto-regulatorio-brasileiro-o-que-bacen-cvm-anbima-e-lgpd-exigem-antes-do-primeiro-prompt/Thu, 18 Jun 2026 12:00:00 GMTPart 2 of 6 in the series Generative AI in a Regulated Financial Institution.https://brunovicco.github.io/technical-articles/en/articles/ia-generativa-em-ambientes-regulados/https://brunovicco.github.io/technical-articles/en/articles/ia-generativa-em-ambientes-regulados/Tue, 16 Jun 2026 12:00:00 GMTIn a regulated financial institution, the criterion is not speed: it is demonstrability. This introduction establishes the thesis, the Product ≠ Provider principle, and the map of the next five parts.https://brunovicco.github.io/technical-articles/en/articles/o-pipeline-completo-do-texto-do-usuario-a-resposta-formatada/https://brunovicco.github.io/technical-articles/en/articles/o-pipeline-completo-do-texto-do-usuario-a-resposta-formatada/Tue, 28 Apr 2026 12:00:00 GMTOver the past few months, I have published separate posts about each component of the e-commerce conversational assistant: episodic memory, routers, DSPy, anaphora, Redis Stack, guardrails, observability, and QueryBuilder.https://brunovicco.github.io/technical-articles/en/articles/seven-frameworks-for-governing-llm-systems-the-map-before-the-territory/https://brunovicco.github.io/technical-articles/en/articles/seven-frameworks-for-governing-llm-systems-the-map-before-the-territory/Tue, 28 Apr 2026 12:00:00 GMTI've been working for twenty years in banking. One thing that this time taught me is that compliance and security frameworks don't exist for bureaucratic reasons, they exist because someone, at some point, paid a steep price for not…https://brunovicco.github.io/technical-articles/en/articles/querybuilder-como-transformar-um-objeto-pydantic-numa-query-ft-search-segura/https://brunovicco.github.io/technical-articles/en/articles/querybuilder-como-transformar-um-objeto-pydantic-numa-query-ft-search-segura/Mon, 20 Apr 2026 12:00:00 GMTIn the posts about Redis Stack, I showed the Fat/Slim/Hybrid model and how the modeling decision directly affects what the LLM receives. Today I want to show the layer between DSPy's RouterOutput and Redis: the QueryBuilder.https://brunovicco.github.io/technical-articles/en/articles/dspy-refine-autocorrecao-em-runtime-sem-recompilar-o-modelo/https://brunovicco.github.io/technical-articles/en/articles/dspy-refine-autocorrecao-em-runtime-sem-recompilar-o-modelo/Mon, 13 Apr 2026 12:00:00 GMTIn the posts about DSPy, I showed how BootstrapFewShot compiles a router that classifies intent and extracts structured parameters. That is DSPy in offline mode: you compile once, version the JSON, and use it in production.https://brunovicco.github.io/technical-articles/en/articles/observabilidade-num-grafo-langgraph-o-que-o-langfuse-ve-que-o-log-nao-ve/https://brunovicco.github.io/technical-articles/en/articles/observabilidade-num-grafo-langgraph-o-que-o-langfuse-ve-que-o-log-nao-ve/Wed, 08 Apr 2026 12:00:00 GMTIn previous posts, I showed how DSPy classifies intent and how the custom embedding-based router decides the macro-intent. The two have one thing in common: they can degrade silently.https://brunovicco.github.io/technical-articles/en/articles/tres-routers-tres-problemas-diferentes-dspy-semantic-router-customizado-e-aurelio-ai/https://brunovicco.github.io/technical-articles/en/articles/tres-routers-tres-problemas-diferentes-dspy-semantic-router-customizado-e-aurelio-ai/Wed, 01 Apr 2026 12:00:00 GMTIn previous posts I showed how DSPy replaced Agno/LangGraph as an assistant's cognitive router, adding an output contract, a metric, and an improvement process on top of the LLM.https://brunovicco.github.io/technical-articles/en/articles/dspy-na-pratica-o-que-muda-quando-o-roteador-ja-e-llm-mas-ainda-nao-e-compilavel/https://brunovicco.github.io/technical-articles/en/articles/dspy-na-pratica-o-que-muda-quando-o-roteador-ja-e-llm-mas-ainda-nao-e-compilavel/Thu, 26 Mar 2026 12:00:00 GMTIn the latest posts, I discussed data modeling in Redis Stack and DSPy as a framework that treats prompts as compilable code.https://brunovicco.github.io/technical-articles/en/articles/guardrails-regulatorios-em-assistentes-de-investimentos-cvm-anbima-e-o-paradoxo-da-lgpd/https://brunovicco.github.io/technical-articles/en/articles/guardrails-regulatorios-em-assistentes-de-investimentos-cvm-anbima-e-o-paradoxo-da-lgpd/Sun, 22 Mar 2026 12:00:00 GMTThe other day, I discussed the taxonomy of memory (episodic, semantic, and procedural) and how each type requires a different infrastructure layer in a financial assistant.https://brunovicco.github.io/technical-articles/en/articles/memoria-em-agentes-episodica-semantica-e-procedural/https://brunovicco.github.io/technical-articles/en/articles/memoria-em-agentes-episodica-semantica-e-procedural/Mon, 16 Mar 2026 12:00:00 GMTWhen you build a conversational assistant for e-commerce or a product catalog, the persistence question is technical: latency, cost, TTL.https://brunovicco.github.io/technical-articles/en/articles/dspy-o-framework-que-trata-prompts-como-codigo-compilavel-nao-como-strings/https://brunovicco.github.io/technical-articles/en/articles/dspy-o-framework-que-trata-prompts-como-codigo-compilavel-nao-como-strings/Sat, 14 Mar 2026 12:00:00 GMTYou know when a model version is about to be deprecated, or you need to switch to another provider's model, and that means adjusting your prompts to make them work? One of DSPy's purposes is to help with that pain point.https://brunovicco.github.io/technical-articles/en/articles/fat-vs-slim-vs-hibrido-no-redis-stack-o-modelo-que-mudou-minha-forma-de-pensar-em-retrieval-para-llm/https://brunovicco.github.io/technical-articles/en/articles/fat-vs-slim-vs-hibrido-no-redis-stack-o-modelo-que-mudou-minha-forma-de-pensar-em-retrieval-para-llm/Thu, 12 Mar 2026 12:00:00 GMT(A continuation of the previous post. If you want the full context, start there.)https://brunovicco.github.io/technical-articles/en/articles/observabilidade-em-ia/https://brunovicco.github.io/technical-articles/en/articles/observabilidade-em-ia/Tue, 25 Mar 2025 12:00:00 GMTImagine flying an airplane without navigation instruments—that is how many teams feel when deploying AI systems into production without a robust observability strategy. And, let's be honest, nobody wants their system to become a ticking…https://brunovicco.github.io/technical-articles/en/articles/arquiteturas-resilientes-em-ia-preparando-se-para-o-inevitavel/https://brunovicco.github.io/technical-articles/en/articles/arquiteturas-resilientes-em-ia-preparando-se-para-o-inevitavel/Wed, 12 Mar 2025 12:00:00 GMTMonths of AI development, thousands invested—and at the most critical moment, your system freezes. This is not a hypothetical scenario. It happens. More often than you might think.https://brunovicco.github.io/technical-articles/en/articles/building-an-end-to-end-ai-governance-framework/https://brunovicco.github.io/technical-articles/en/articles/building-an-end-to-end-ai-governance-framework/Tue, 04 Mar 2025 12:00:00 GMTPart 4 of 4 in the AI Security for Enterprises Serieshttps://brunovicco.github.io/technical-articles/en/articles/proactive-security-and-real-time-threat-detection-in-ai-systems/https://brunovicco.github.io/technical-articles/en/articles/proactive-security-and-real-time-threat-detection-in-ai-systems/Mon, 24 Feb 2025 12:00:00 GMTPart 3 of 4 in the Enterprise AI Security Serieshttps://brunovicco.github.io/technical-articles/en/articles/ai-platform-security-features/https://brunovicco.github.io/technical-articles/en/articles/ai-platform-security-features/Mon, 10 Feb 2025 12:00:00 GMTPart 2 of 4 in the Enterprise AI Security Serieshttps://brunovicco.github.io/technical-articles/en/articles/understanding-ai-security-fundamentals-terms-of-service-and-data-privacy/https://brunovicco.github.io/technical-articles/en/articles/understanding-ai-security-fundamentals-terms-of-service-and-data-privacy/Mon, 03 Feb 2025 12:00:00 GMTPart 1 of 4 in the Enterprise AI Security Serieshttps://brunovicco.github.io/technical-articles/en/articles/from-rag-to-hybrid-architectures/https://brunovicco.github.io/technical-articles/en/articles/from-rag-to-hybrid-architectures/Sun, 02 Feb 2025 12:00:00 GMTThe landscape of enterprise information retrieval is rapidly evolving beyond basic RAG implementations. While the foundational RAG approach revolutionized how AI systems access and utilize information, today's enterprise needs are…https://brunovicco.github.io/technical-articles/en/articles/choosing-the-right-approach-rag-prompt-engineering-or-fine-tuning/https://brunovicco.github.io/technical-articles/en/articles/choosing-the-right-approach-rag-prompt-engineering-or-fine-tuning/Tue, 28 Jan 2025 12:00:00 GMTIn the landscape of generative AI and Large Language Models (LLMs), choosing the right technique for solving a particular problem can feel overwhelming. RAG, Prompt Engineering, and Fine-Tuning are three powerful tools in your AI…https://brunovicco.github.io/technical-articles/en/articles/fine-tuning-llms-a-deep-dive-into-instruction-fine-tuning/https://brunovicco.github.io/technical-articles/en/articles/fine-tuning-llms-a-deep-dive-into-instruction-fine-tuning/Wed, 22 Jan 2025 12:00:00 GMTAs we navigate the evolving landscape of language models, fine-tuning has emerged as a crucial technique for adapting these powerful tools to specific needs. Through my experience focusing on instruction fine-tuning, combined with…https://brunovicco.github.io/technical-articles/en/articles/the-art-and-science-of-prompt-engineering-unlocking-llm-potential/https://brunovicco.github.io/technical-articles/en/articles/the-art-and-science-of-prompt-engineering-unlocking-llm-potential/Mon, 20 Jan 2025 12:00:00 GMTAs Large Language Models (LLMs) continue to reshape our technological landscape, the role of prompt engineering has emerged as a crucial discipline at the intersection of human communication and artificial intelligence. In the past…https://brunovicco.github.io/technical-articles/en/articles/from-search-to-generation-unlocking-the-power-of-retrieval-augmented-generation-rag/https://brunovicco.github.io/technical-articles/en/articles/from-search-to-generation-unlocking-the-power-of-retrieval-augmented-generation-rag/Tue, 14 Jan 2025 12:00:00 GMTImagine you're chatting with an AI assistant. You ask it a complex question, and rather than giving you vague, generic advice, it offers a detailed, accurate response grounded in real-world facts. The secret behind such a feat?…https://brunovicco.github.io/technical-articles/en/articles/from-theory-to-practice-building-enterprise-ai-solutions-with-teams-integration/https://brunovicco.github.io/technical-articles/en/articles/from-theory-to-practice-building-enterprise-ai-solutions-with-teams-integration/Sun, 12 Jan 2025 12:00:00 GMTAfter dedicating my efforts to developing and implementing AI solutions through Microsoft Teams, I've discovered that the true power of AI lies not in complex algorithms, but in practical, day-to-day applications that transform how…https://brunovicco.github.io/technical-articles/en/articles/aiopslab-automacao-de-operacoes-em-nuvem/https://brunovicco.github.io/technical-articles/en/articles/aiopslab-automacao-de-operacoes-em-nuvem/Mon, 30 Dec 2024 12:00:00 GMTMicrosoft Research has just introduced a significant innovation in cloud operations: AIOpsLab, a comprehensive framework for developing and evaluating AI agents for autonomous cloud operations.https://brunovicco.github.io/technical-articles/en/articles/meta-apresenta-os-large-concept-models/https://brunovicco.github.io/technical-articles/en/articles/meta-apresenta-os-large-concept-models/Sun, 29 Dec 2024 12:00:00 GMTThe FAIR (Meta AI Research) team has just published research that redefines our understanding of language models. The new Large Concept Model (LCM) introduces a fundamentally different approach to how AI processes and generates content.https://brunovicco.github.io/technical-articles/en/articles/summary-of-ai-regulations-eu-and-usa/https://brunovicco.github.io/technical-articles/en/articles/summary-of-ai-regulations-eu-and-usa/Mon, 07 Oct 2024 12:00:00 GMTLast week, we explored the AI regulatory landscapes in the European Union and the United States. Today, let's dive deeper and compare these two approaches, highlighting the key aspects that set them apart.https://brunovicco.github.io/technical-articles/en/articles/ai-regulation-in-the-usa-a-patchwork-approach/https://brunovicco.github.io/technical-articles/en/articles/ai-regulation-in-the-usa-a-patchwork-approach/Thu, 03 Oct 2024 12:00:00 GMTWhile the EU moves forward with its comprehensive AI Act, the United States has taken a different path in regulating artificial intelligence. Instead of a single, overarching framework, the US approach to AI regulation is characterized…https://brunovicco.github.io/technical-articles/en/articles/the-eu-artificial-intelligence-act-shaping-the-future-of-ai-regulation/https://brunovicco.github.io/technical-articles/en/articles/the-eu-artificial-intelligence-act-shaping-the-future-of-ai-regulation/Mon, 30 Sep 2024 12:00:00 GMTIn an era where artificial intelligence (AI) is rapidly transforming industries and societies, the European Union has taken a bold step forward with the proposed Artificial Intelligence Act (AI Act). This groundbreaking legislation aims…