Rogerio Rodrigues da Silva

I was revisiting some of my early articles on crypto and stumbled upon this one, published back on January 23, 2020: https://lnkd.in/dcHwBcBV (in Portuguese). In it, I used a simple bootstrap simulation to generate return scenarios across different holding periods. The goal was to argue that crypto is better suited for long-term investments—showing that the longer the holding period, the less the outcome resembled a coin toss. This table below was extracted from the article and summarizes its results. I remember receiving feedback like: “Interesting exercise, but you’re clearly overestimating the odds of very high returns. A 10x in 5 years is highly unlikely, and you’re assigning it a 63% probability.” To be honest, I had my own doubts about those figures. I never took them at face value, given the oversimplified method used. Still, according to Yahoo Finance, Bitcoin was priced at $8,406.52 on the day the article was published. Five years later, it reached $103,960.17—a 12.4x increase. Sure, it was a lucky shot from the simplistic method I built. But the point of this post isn’t about statistical models—it's about how #crypto constantly forces us to recalibrate our mental models of what’s likely or not. Sometimes for worse, but most often for better outcomes.

51

1 comentário

AWS re:Invent: “Artificial intelligence is no longer an experiment — it is becoming the foundation of how the world builds, scales, and innovates.” Grande evento da Amazon Web Services (AWS) inovando e direcionando o mercado de tecnologia. Alguns insights: 1. The Dawn of the Renaissance Developer: … is coriuos … thinks in systems … communicates … is an owner … is a polymath 2. AI tools redefining how we build, deploy, and transform software: • AWS Kiro — turning ideas into production-ready architectures • AWS DevOps Agent — automating delivery with intelligence and speed • AWS Transform — modernizing legacy into cloud-native reality 3. Payments • AI-native fraud prevention — moving from reactive controls to real-time, predictive risk decisions • Instant, always-on payments — architectures designed for low latency, high availability, and global reach • Modernization at scale — transforming legacy payment systems into cloud-native, event-driven platforms • Embedded intelligence — AI infused across authorization, reconciliation, dispute handling, and customer experience • Developer velocity as a differentiator — automation and AI agents accelerating delivery, resilience, and compliance Agora o desafio é fazer acontecer.

197

I recently came across this video (Attekita Dev is a great channel, by the way) about building a software engineering portfolio: https://lnkd.in/dijnsm6S If you’re just starting your developer career, a portfolio can be a real game changer. It shows what you can do even when you don’t yet have much formal experience. It helps you stand out from other beginners and gives you real stories to share in interviews, boosting your confidence. I also wrote more about this here: https://lnkd.in/dP5hx8XW To get started: * Choose one small project, even a simple one from a course, and make it yours. * Document not just what you built, but why you built it and what you learned. * Keep improving it over time, growth is more impressive than perfection.

25

2 comentários

Hello World!! Our article is now live on the AWS Machine Learning Blog, co-written by Instituto de Ciência e Tecnologia Itaú - ICTi and Amazon Web Services (AWS), on sentiment analysis across text and speech using AWS Generative AI services. We evaluated two practical approaches to extract “voice of the customer” signals at scale: 1) Audio → transcription → text sentiment (Amazon Transcribe + Amazon Comprehend / FMs via Amazon Bedrock & Amazon SageMaker) 2) Direct sentiment from the audio signal, preserving prosody/intonation (HuBERT, Wav2Vec, Whisper on Amazon SageMaker) Key takeaways: text-only sentiment can fail on subtle cases (sarcasm/ambiguity). Audio-native models capture emotional cues better in controlled settings—but still struggle to generalize across more diverse phrases and contexts, reinforcing the need for multimodal strategies (text + audio). We also discuss how to orchestrate these components on AWS (for example, Kinesis/Lambda/S3 and integration with Amazon Connect) to support near-real-time analytics. 🔗 https://lnkd.in/dWq4QbD9 Authors: Caique de Almeida; Guilherme Rinaldo, Paulo Finardi, Victor Costa Beraldo, Vinicius Caridá, Ph.D. and Pranav Murthy #AWS #GenerativeAI #MachineLearning #NLP #SpeechAI #SentimentAnalysis #CustomerExperience #ContactCenter #MLOps Itaú Unibanco

97

2 comentários

Paper recomendo a leitura se tiver interesse em entender uma das limitações dos LLMs no momento e o que a Bytedance sugere em parecia com universidade alternativa. Vai ser ótimo quando ver parcerias com essa por aqui. "Large language models (LLMs) often fail to learn effective long chain-of-thought (Long CoT) reasoning from human or non-Long-CoT LLMs imitation. To understand this, we propose that effective and learnable Long CoT trajectories feature stable molecular-like structures in unified view, which are formed by three interaction types: Deep-Reasoning (covalent-like), Self-Reflection (hydrogen-bond-like), and Self-Exploration (van der Waals-like). Analysis of distilled trajectories reveals these structures emerge from Long CoT fine-tuning, not keyword imitation. We introduce Effective Semantic Isomers and show that only bonds promoting fast entropy convergence support stable Long CoT learning, while structural competition impairs training. Drawing on these findings, we present Mole-Syn, a distribution-transfer-graph method that guides synthesis of effective Long CoT structures, boosting performance and RL stability across benchmarks." 2601.06002v2.pdf https://lnkd.in/dcThHdPh

15

🔐 Beyond HTTPS — How I Secured .NET APIs for 40,000+ Users with Azion Let me tell you — securing APIs is more than just adding HTTPS. It’s about protecting your users, your data, and your business. In one of my cloud projects, I built an API that serves over 40,000 users, and I made sure it was fast, secure, and resilient. How? With a collaborative approach, working alongside the Azion team and the client’s infrastructure team. ⸻ ✅ 1. Authentication & Authorization (API-Level Security) • 🔑 JWT (JSON Web Token): Secure, stateless authentication for APIs. • 🔒 OAuth 2.0 + Microsoft Identity: For secure enterprise login. • 📝 Role-Based (RBAC) + Policy-Based Authorization: Fine-grained access control. • 🔎 Claims-Based Access: Dynamic permissions based on user data. ⸻ ✅ 2. Edge Security with Azion (Protecting 40,000+ Users) But we didn’t stop there — together with the Azion team and the client’s infrastructure team, we added: • 🛡️ Edge Firewall: Blocking malicious traffic (DDoS protection, IP restrictions). • 🔎 WAF (Web Application Firewall): Protecting against OWASP Top 10 threats (SQL Injection, XSS). • 🔁 Reverse Proxy: Securely routing traffic to the APIs, hiding their real IPs. • 🔒 SSL Offload: Azion managed SSL/TLS, so I didn’t have to worry about certificates. • 🚀 Edge Caching: Speeding up static content delivery (CDN). • 📊 Real-Time Monitoring: Detecting and tracking traffic anomalies. ⸻ ✅ 3. Monitoring & Visibility Because security without visibility is just hope: • 📊 Application Insights (Azure): Complete visibility on API performance and errors. • 📝 Structured Logging (Serilog): JSON logs with full context (user, request, response). • 🔎 Real-Time Alerts: Detecting unusual behavior and notifying instantly. ⸻ 💡 Real-World Impact: • Secured over 40,000 users with JWT and Microsoft Identity. • All traffic passed through Azion, providing: • 🔒 SSL offload (Azion managed HTTPS). • 🛡️ WAF (blocking malicious requests at the edge). • 🔁 Reverse Proxy (hiding API endpoints). • Static content was cached at the edge, improving performance. • Monitoring with Application Insights for complete visibility. ⸻ 🚀 Why this worked: • ✅ Secure: Only authorized users access the API. • ✅ Fast: SSL offload + CDN with Azion. • ✅ Resilient: Attacks are blocked at the edge. • ✅ Scalable: 40,000+ users with zero downtime. • ✅ Visible: Full monitoring with Application Insights. ⸻ ✨ This wasn’t just my work — it was a collaborative success, with expertise from the Azion team and the client’s infrastructure team. 👥 Follow me for more real-world .NET and cloud solutions. #dotnet #api #security #cloudcomputing #jwt #azure #applicationinsights #azion #reallifecode #devlife #linkedinbr

29

5 comentários

🇧🇷 "Stablecoins shouldn’t be treated like traditional FX. The future of global payments needs a new playbook." The future of global payments is increasingly shaped by stablecoins, local payment systems like Pix, and new open financial infrastructure. In this recent interview with Cointelegraph Brasil , Nkiru Uwaje shared some of the challenges and opportunities we are seeing in this space: 👉 “Brazil wants to protect currency controls and data sovereignty, but classifying stablecoins as foreign exchange operations creates silos and operational complexity for companies.” At MANSA, we are building towards a global layer of liquidity and interoperability connecting stablecoins with local payment systems like Pix, in a more efficient and secure way. We are very proud to contribute to this important conversation in Brazil — a market with huge potential to lead this transformation globally. Full article here: 👉 https://lnkd.in/dxqWxveY #Mansa #Stablecoins #CrossBorder #Pix #FutureOfPayments #Web3 #Fintech #Brazil

23

Node.js has always fascinated me, not just for what it does, but how it does it. It’s built on a simple idea => keep the core minimal and build everything else as small, independent modules. That mindset is what makes it so efficient, and so uniquely “Node.” I recently picked up "Node.js Design Patterns (4th Edition)" by 🦁 Luciano Mammino and Mario Casciaro, and it reminded me just how deep the platform really goes. 📗 The Book - https://lnkd.in/dGcr8xw2 They don’t just explain patterns, they connect them to how Node.js actually works: ▣ How the Reactor pattern powers everything under the hood ▣ Why async design isn’t just about callbacks, but architecture ▣ Streams as a fundamental design primitive ▣ The evolution from callbacks → promises → async/await ▣ Real-world DI, factory, and middleware patterns that actually fit Node.js ▣ Scaling and messaging patterns that make systems production-ready I genuinely enjoyed the book because it made the connection between the theory of the event loop and the practical design of a large application feel direct and obvious. Highly recommended.

314

77 comentários

Getnet at the forefront of Agentic Commerce, unveiling its strategy at NRF 2026 event in New York 🤖 Commerce is entering a new era. #AI is no longer limited to analysing customer journeys, it is becoming an active force in how commerce happens. As autonomous agents discover products, make decisions and execute #payments, the role of payments is evolving from a final transactional step into a strategic layer of trust, control and value creation. In this context, last week, alongside #NRF2026 in New York, Getnet presented its #AgenticCommerce strategy at an exclusive event for clients and partners. This marks a significant milestone for Getnet, outlining a comprehensive, research-backed approach to Agentic Commerce for #LatinAmerica and #Iberia, developed in collaboration with Monitor Deloitte. According to market estimates, Agentic AI is expected to influence up to 30% of global e-commerce transaction value by 2030, reinforcing the need for payment infrastructures prepared to securely support transactions initiated by intelligent agents. As Juan Franco, CEO of Getnet, explains: “Agentic Commerce marks the moment when AI stops observing the customer journey and starts driving it. Our vision is to be our customers’ trusted advisor and enabler, transforming our acquisition, risk management and data capabilities into an open, intelligent platform that allows merchants to handle AI agents as easily as they handle human customers today.” At #Getnet, innovation is how we help define what comes next. We empower our customers to build trust at scale, unlock intelligence across every interaction and lead the transformation of commerce in an increasingly autonomous, omnichannel world. 📘 Read the full whitepaper and explore Getnet’s Agentic Commerce strategy: https://lnkd.in/edMzpVhD

201

3 comentários

As novidades não param, veja esse novo protocolo AP2: Powering AI commerce with the new Agent Payments Protocol (AP2) Today, Google announced the Agent Payments Protocol (AP2), an open protocol developed with leading payments and technology companies to securely initiate and transact agent-led payments across platforms. The protocol can be used as an extension of the Agent2Agent (A2A) protocol and Model Context Protocol (MCP). In concert with industry rules and standards, it establishes a payment-agnostic framework for users, merchants, and payments providers to transact with confidence across all types of payment methods. We’re collaborating with a diverse group of more than 60 organizations to help shape the future of agentic payments, including Adyen, American Express, Ant International, Coinbase, Etsy, Forter, Intuit, JCB, Mastercard, Mysten Labs, Paypal, Revolut, Salesforce, ServiceNow, UnionPay International, Worldpay, and more. https://lnkd.in/dNVtR6xz

27

Quer ver os commits do Git, por autor? Esse é um alias legal: commits-by-author = "!f() { git log --author=\"$1\" --name-status --pretty=format:\"%C(yellow)%h%Creset %C(green)%ad%Creset %C(blue)%an%Creset: %s\" --date=short -m; }; f" Pode ser utilizado assim: $ git commits-by-author "Eustaquio Rangel" 3c08518 2025-02-26 Eustaquio Rangel: Showing constructors on method listing M Gemfile.lock M lib/jsearch/searcher.rb M lib/jsearch/version.rb

22

2 comentários