Codekerdos’ Post

Excited to share that I attended the 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜 𝗳𝗼𝗿 𝗗𝗲𝘃𝗢𝗽𝘀 𝗠𝗮𝘀𝘁𝗲𝗿𝗰𝗹𝗮𝘀𝘀 by Shubham Londhe yesterday! 🚀 The session gave me a completely new perspective on how AI is transforming modern DevOps workflows. We are moving beyond traditional automation into systems that can analyze, reason, and take intelligent actions autonomously. Some of the key takeaways from the masterclass: ✅ Understanding how AI agents can automate DevOps decision-making ✅ Learning real-world use cases of AI in cloud and infrastructure management ✅ Exploring LLM-driven monitoring and troubleshooting workflows ✅ Hands-on exposure to integrating AI with Kubernetes and AWS environments ✅ Discovering how Agentic AI can improve operational efficiency and incident handling This hands-on learning experience reinforced one thing clearly: 𝗔𝗜 + 𝗗𝗲𝘃𝗢𝗽𝘀 𝗶𝘀 𝗴𝗼𝗶𝗻𝗴 𝘁𝗼 𝗿𝗲𝗱𝗲𝗳𝗶𝗻𝗲 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 𝗺𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁. Big thanks to TrainWithShubham for such an insightful and practical session. Looking forward to exploring more in the world of Agentic AI and AIOps. #DevOps #AgenticAI #AIOps #AWS #Kubernetes #CloudComputing #Python #LLM #InfrastructureAutomation #Learning

Debugging Kubernetes issues shouldn’t take hours. With Kubegraf, you can detect problems, identify the root cause, and apply fixes in seconds. No more digging through endless logs or guessing what went wrong. Focus on building, not debugging. Try now :- www.kubegraf.io #Kubernetes #DevOps #SRE #CloudComputing #AI #PlatformEngineering #Observability #CNCF

This is the AI productivity trap. The tools are generating more code, faster — but without the right guardrails, reviews, and delivery infrastructure underneath, you're accelerating toward technical debt, not away from it. And it gets worse. 42% say the cost of AI tools is becoming a real problem. 31% say tool proliferation is making it impossible to choose wisely. More tools. More code. Less clarity. More cost. This is exactly the world Opsera was built for. When your engineers are running Claude Code, Gemini Code Assist, GitHub Copilot, and three other AI tools across different teams and pipelines, someone needs a unified view of what's actually happening — what's shipping, what's working, and what's quietly accumulating risk. Speed without visibility isn't DevOps. It's just chaos at a higher velocity. #DevOps #AI #Opsera #EngineeringLeadership #SoftwareDelivery #PlatformEngineering #AIGovernance #CICD

- Building a production-ready AI system is less about the "model" and more about the orchestration. Lately, I’ve been focusing on moving beyond simple scripts toward a fully decoupled, containerized architecture. The goal is to ensure every component, from real-time data ingestion to the final user interface, is scalable and robust. By isolating tasks like document-based knowledge querying and deep learning model training into specialized Docker containers, we gain the flexibility to iterate on AI logic without disrupting core data pipelines. It’s been rewarding to manage the full ML lifecycle from architecture design to deployment to solve complex industrial challenges. - How are you handling the complexity of modern MLOps? I'd love to hear your thoughts! #SoftwareArchitecture #GenAI #Microservices #MLOps #Docker #LangChain #FullStack #AIEngineering #NorwayTech

DevOps Concept of the Day: Automated Model Retraining Automated retraining closes the ML loop: monitor → detect drift → retrain → evaluate → deploy if better. Champion-challenger pattern tests new vs current model before any swap. Full automation of the ML lifecycle. Today's DevOps/MLOps update (MLflow): TypeScript SDK 0.2.0 Bump several RC TypeScript packages stable version. https://lnkd.in/dGySBF4g Why it matters: MLOps closes the gap between data science research and production reliability. #MLOps #AutomatedML #MachineLearning #AI

🤖 Docker Meets AI with Gordon 🚀 Was exploring Gordon AI inside Docker, Inc Desktop today, and it actually feels super useful for developers 👀 🐳 Helps write & optimize Dockerfiles 🛠️ Assists in debugging containers 📦 Simplifies Docker Compose setups 🔐 Scans images for vulnerabilities ⚡ Explains logs & runtime issues What I liked most? Everything works directly inside Docker Desktop, so less tab switching and faster workflows 🔥 #Docker #DevOps #AI #SoftwareEngineering

🧠 One of the strongest ideas in AI-DLC is that context should get richer, not weaker, as you move forward. That is something I think many AI coding experiences still get wrong. They can be useful at the beginning, but they often lose coherence as the project grows. AI-DLC solves that by making each phase feed the next one with more structured artifacts. -> In the inception phase, you clarify intent. -> In construction, you model the system and generate implementation. -> In operations, you start thinking about deployment, observability, and continuous validation. That progression is powerful because it turns AI from a one-shot generator into a lifecycle participant. Kiro is interesting here because it helps preserve that structure through specs, steering files, and hooks. That means you are not just asking AI to write code — you are asking it to work inside a controlled context that reflects the current stage of the project. For people building AWS solutions, this matters a lot. The difference between a demo and something actually scalable is often the ability to keep architecture, implementation, and delivery aligned. That is where Kiro and AI-DLC start to make more sense together. #AWS #Kiro #AIDLC #SpecDrivenDevelopment #AIEngineering #CloudComputing #SoftwareEngineering #DevOps #GenAI #DevEx

Just completed Agent Skills and Subagents from Anthropic Academy — and if you've been following my AI learning journey, this is where things get genuinely complex. Here's the core idea: instead of one AI doing everything, you break the work into specialized agents. One plans, one executes, one verifies. Each subagent handles a focused task, and together they tackle problems that a single prompt never could. Think of it like a well-structured engineering team — except every member responds in milliseconds. Why this matters in the real world: 🔗 Long, multi-step workflows can now be automated end-to-end 🧩 Each subagent stays focused, which means fewer errors and more predictable outputs 🏗️ Complex infrastructure tasks — provisioning, testing, validating — become orchestratable 📋 You stop thinking in prompts and start thinking in systems As someone who designs distributed systems and CI/CD pipelines for a living, this mental model felt immediately familiar. It's basically microservices architecture — but for AI workflows. #AgenticAI #Anthropic #AIAgents #DevOps #CloudEngineering #AWS #SystemDesign #ContinuousLearning #BuildInPublic

It is encouraging to see the industry waking up to the reality that AI inference is commoditizing. But there is a crucial part of the story still missing: the data layer. Great to see Devansh Saxena breaking down the fundamental difference between a static context graph and a stateful context graph below (see comments). In simple terms: it is the difference between a standard Google Map and Google Maps with real-time traffic data. Which one actually gets you to resolution faster during a production outage? Once the market fully understands that distinction, the next logical question becomes: what is the true acquisition cost to achieve it? https://lnkd.in/g4BA6Ru9

#AI writes your #code. Cool. But who writes your requirements? Your architecture? Your deployment strategy? That's the gap nobody's talking about. AWS GenAI Virtual Tech Track is back! On June 2 we're going deep on the AI Development Life Cycle (AI-DLC) - a structured methodology that puts AI at the center of the entire #SDLC. Not just autocomplete. The full thing. What you'll get: → Where Coding AI actually adds value beyond code generation → How to measure developer productivity in an AI-assisted world → Real-world examples of AI-DLC in practice with Kiro Join Dror Helper and Dana Mordechai Ganzer for a 2-hour deep dive. June 2 | 11:00 AM IDT | Online Zoom Webinar - in Hebrew Register: https://lnkd.in/ezqGpwfc #AIDLC #Kiro #GenAI #AWS #DeveloperProductivity #SoftwareEngineering