Mohammad Reza Moattar

We’re moving into an era where autonomous software agents will inherently change governance, work, and societal resilience. But how do we succeed at it? Building open, sovereign, and trustworthy digital ecosystems based on open data, source, standards, and skills. Developing platforms that allow shared skills and capabilities across institutions, making sure that AI serves the public good while preserving sovereignty. We are now reimagining how we govern, collaborate, and build a digital society rooted in transparency, accountability, and shared responsibility. A hearty thank you to tahawul tech and Sandhya D'Mello for featuring my thoughts on building an agentic future that is open by design. Read the full article here:  https://lnkd.in/enfMz_dA

158

Most execs want to know: “What AI tools should we use?” Wrong question. Ask this instead: “Where in our workflows is value created—or lost?” 👇 That’s where AI belongs. Three months ago, I helped a mid-size services firm stop chasing 7 disconnected pilots. We started with a simple map: • 🧭 Value stages • 🔁 Workflow loops • ⚠️ Friction zones Then we scored each one: ✅ High margin impact ⚙️ Manual decision + data-heavy 🔁 Repeated 10+ times a day Guess what? Only 2 workflows passed the test. That’s where we decided to deploy AI. And yes—results are in. AI doesn’t scale by stacking tools. It scales by unlocking bottlenecks. 💥 Tip: You don’t need an AI strategy. You need a business strategy that knows where AI fits. 🧭 How to start: • Template your firm (and industry) value chain • Break it into workflows and tasks • Find the friction—delay, manual effort, complexity • Score each area for value, frequency, and AI fit • Start small, scale fast 📩 Leading an AI Transformation? Download the Cheatsheet, Connect. ♻️ Repost if you’re done confusing AI with tools. ➕ Follow Adi Agrawal | Real Leadership

67

70 Comments

Building Event-Driven Microservices: What an engaging August session at the Data Book Club! Led by Michael Pike, we dove into Adam Bellemare "Building Event-Driven Microservices" and sparked some fantastic discussions around the future of real-time data and scalable systems.   Key Insights 🔺 Event-Driven Microservices: We explored how shifting from traditional batch processing to event-driven architectures allows for real-time data sharing and truly scalable solutions. ▪️Data Liberation: A standout takeaway was the concept of extracting data from legacy systems and publishing it into event streams—unlocking data for modern architectures without major upheaval. We examined query-based, log-based, and table-based extraction methods, weighing their suitability for different platforms. ▪️Message Brokers vs. Event Brokers: a clear distinction between these two, focusing on the advantages of retaining and replaying event streams for resilient, flexible data pipelines. ▪️Workflow Patterns: Choreography vs. orchestration—understanding when to use each for managing workflows in microservices, and the ramifications for scalability, complexity, and error handling. ▪️Architecture Tools: We discussed the landscape of containers, orchestration frameworks (like Kubernetes, Kafka, Spark, Flink), and the so-called “microservice tax”—the costs and considerations of adopting event-driven infrastructure. The conversation touched on both heavyweight and lightweight orchestration models for optimal scalability and state management. ▪️Real-World Application: The club agreed: it's time to double down on challenging conventional batch data flows and embrace event streaming as a foundation for both new and existing platforms. 🔺Looking Forward: As AI becomes increasingly central, event-driven architectures are set to play a crucial role in supporting intelligent, adaptive systems.   💡 Main advice from the book: Consider how you can move from batch to real-time in your own initiatives. Liberate your data, rethink your architectures, and help build the future of scalable, event-driven systems!   Check out the book for a whole lot more insight. Our Westpac data book club is our member driven community for our curious & continuous learners, data and tech enthusiasts.   #databookclub #eventdrivenmicroservices #eventdriven #microservices #realtimedata Thomas Bauch Gerard Seeto

41

2 Comments

🚀 Platform-Agnostic Deployments with ArgoCD – GitOps in Action One of the biggest challenges in today’s hybrid and multi-cloud environments is ensuring consistent deployments without being locked into a single platform. That’s where ArgoCD shines. 🔑 With its GitOps-first approach, ArgoCD enables a “single source of truth” for deployments—whether on AWS, Azure, GCP, or on-prem. Here’s why enterprises are embracing ArgoCD for platform-agnostic CD pipelines: ✅ Multi-cluster & Multi-cloud: Deploy to Kubernetes clusters across providers with the same pipeline ✅ Declarative GitOps model: Version-controlled manifests ensure transparency and rollback ✅ Automated sync: Continuous reconciliation between Git state and cluster state ✅ Scalable & Secure: Role-based access, SSO integration, and audit trails baked in ✅ Extensible: Integrates with Terraform, Helm, Kustomize, and service meshes 🌎 The outcome? Build once, deploy anywhere → without re-engineering Compliance by design → every change tracked in Git Developer velocity → self-service deployments with guardrails 💡 In the era of multi-cloud agility, ArgoCD is more than a tool—it’s the foundation of a modern, platform-agnostic CD strategy.

46

A Post-Hype view of AI pair programming. AI coding tools are powerful for certain tasks with significant caveats. They can enhance developer learning and enjoyment . The human developer’s role is still central. Understanding the problem, ensuring quality, and making architectural decisions remain human responsibilities that AI can’t shoulder. The developers who benefit most from AI are those who approach it with eyes open, ready to delete most of what it generates if needed, prepared to invest effort in guidance and verification, and determined to use AI to augment good engineering practices, not shortcut them. It can speed up repetitive work and provide inspiration, and it can make coding feel more enjoyable by offloading drudge. But it doesn’t fundamentally change the nature of software development as a thinking intensive, design centric activity. Writing code was never the true bottleneck; understanding requirements, devising solutions, and maintaining quality are where the real effort lies, and those are the aspects where human skill remains irreplaceable.

103

4 Comments

The pendulum has finally begun to swing towards its other extreme: - cloud repatriation vs cloud expansion - systems programming vs web programming - hardware is hot again - open ecosystems vs walled gardens - profit and healthy balance sheets vs  undisciplined growth - microservices back to monoliths - systems administration is exciting again - RISC is taking over  Technology is dynamic and sane again. 

50

2 Comments

Good talk by Rod Johnson. While Python has dominated the Gen AI space, the Java tooling has made significant gains in the last 12 months. His Embabel framework is a very elegant framework for developing agentic solutions for the JVM ecosystem. As this framework matures, I expect to see wide adoption in the enterprise. https://lnkd.in/epZPv8DM

39

Most devs are still building AI wrong. They keep prompting harder when the real unlock is contracts. I just watched the CodeTV × Postman 4-hour challenge, and it flipped my workflow upside down. Here’s the exact playbook that separates weekend hacks from shippable features: Phase 1️⃣: The Proving Ground • Teams had 4 hours to build AI apps from scratch. • Under pressure, bad workflows collapsed. • The good ones? They all had one thing in common. Phase 2️⃣: The Mindset Shift • Stop treating the LLM like a magic generator. • Start treating it like an orchestrator that needs a formal API contract. • Don’t prompt harder. Build smarter contracts. Phase 3️⃣: The Contract Toolkit (Postman MCP) • MCP Generator → wrap any API into an AI-ready tool • MCP Client → test those tools in isolation before the LLM ever touches them This kills the “is it me or the AI?” debugging spiral. Phase 4️⃣: The Winning Workflow 1./ Define the AI’s “toolbelt” with the Generator. 2./ Rigorously test that toolbelt with the Client. 3./ Only then give the LLM the pre-vetted tools. Now you’re not debugging a black box. You’re orchestrating a predictable system. Phase 5️⃣: The New Bottleneck Once this workflow clicks, you see the truth: The hard part isn’t code anymore. It’s orchestration → latency, timeouts, service interaction. This is where the real engineering begins. The result? ✔ 80% less debugging ✔ Faster, more predictable feature delivery ✔ AI features that are reliable, not just clever demos Most devs will keep throwing prompts at the AI. The top 10% will hand it a well-tested API contract. That’s the difference between hacks and production-ready systems. See it in action (Check comments for the Video)

93

35 Comments

✨ 𝗙𝗿𝗼𝗺 𝗔𝗜 𝘁𝗼 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁𝘀 We have moved from 𝘶𝘴𝘪𝘯𝘨 𝗔𝗜 𝗮𝘀 𝗮 𝘁𝗼𝗼𝗹 𝘁𝗵𝗮𝘁 𝘀𝗶𝗺𝗽𝗹𝘆 𝘀𝘂𝗽𝗽𝗼𝗿𝘁𝘀 𝗼𝘂𝗿 𝘄𝗼𝗿𝗸 to 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁𝘀 𝘁𝗵𝗮𝘁 𝗮𝗰𝘁 𝗼𝗻 𝗼𝘂𝗿 𝗯𝗲𝗵𝗮𝗹𝗳. These systems can 𝗽𝗲𝗿𝗰𝗲𝗶𝘃𝗲, 𝗱𝗲𝗰𝗶𝗱𝗲, 𝗮𝗻𝗱 𝘁𝗮𝗸𝗲 𝗮𝗰𝘁𝗶𝗼𝗻 with minimal human oversight. The real power is not just automation, but 𝗮𝘂𝘁𝗼𝗻𝗼𝗺𝘆. AI agents now 𝗹𝗶𝗻𝗸 𝘁𝗮𝘀𝗸𝘀, 𝗮𝗱𝗮𝗽𝘁 𝗶𝗻 𝗿𝗲𝗮𝗹 𝘁𝗶𝗺𝗲, 𝗮𝗻𝗱 𝗲𝘃𝗲𝗻 𝗰𝗼𝗹𝗹𝗮𝗯𝗼𝗿𝗮𝘁𝗲 𝘄𝗶𝘁𝗵 𝗼𝗻𝗲 𝗮𝗻𝗼𝘁𝗵𝗲𝗿. This marks a 𝘀𝗲𝗶𝘀𝗺𝗶𝗰 𝘀𝗵𝗶𝗳𝘁 in how we think about work. Instead of spending time on monitoring and maintenance, we can now focus on 𝗷𝘂𝗱𝗴𝗺𝗲𝗻𝘁, 𝗰𝗿𝗲𝗮𝘁𝗶𝘃𝗶𝘁𝘆, 𝗮𝗻𝗱 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝘆, where 𝘩𝘶𝘮𝘢𝘯 𝘪𝘯𝘴𝘪𝘨𝘩𝘵 𝗿𝗲𝗺𝗮𝗶𝗻𝘀 𝗶𝗿𝗿𝗲𝗽𝗹𝗮𝗰𝗲𝗮𝗯𝗹𝗲. This is only the beginning. Curious about where this shift is headed? I explore it in more depth here: 👉 https://lnkd.in/gh2xmp5Y The rise of 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁𝘀 will not only transform how we work, but also 𝗿𝗲𝘀𝗵𝗮𝗽𝗲 𝗹𝗲𝗮𝗱𝗲𝗿𝘀𝗵𝗶𝗽, 𝗿𝗲𝗶𝗻𝘃𝗲𝗻𝘁 𝗯𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗺𝗼𝗱𝗲𝗹𝘀, 𝗮𝗻𝗱 𝗿𝗲𝗱𝗲𝗳𝗶𝗻𝗲 𝘀𝗼𝗰𝗶𝗲𝘁𝘆 in ways we are 𝗷𝘂𝘀𝘁 𝗯𝗲𝗴𝗶𝗻𝗻𝗶𝗻𝗴 𝘁𝗼 𝗶𝗺𝗮𝗴𝗶𝗻𝗲. #AIagents #FutureOfWork #Leadership #Innovation

36

9 Comments

Thrilled to share that my panel session “Scaling Open Source Development in a Bank” has been selected for the upcoming FINOS Open Source in Finance Forum! 🎉 https://lnkd.in/estivseu I’ll be in a conversation with leaders from Fannie Mae, NatWest, and Morgan Stanley on how financial institutions are launching, operationalizing, and scaling open source readiness programs in highly regulated environments. We’ll cover: 🔹 Tools & frameworks for assessing maturity 🔹 Metrics & KPIs to guide evolution 🔹 Recognition models for contributors 🔹 Governance that scales globally 🔹 Building and retaining open source-aligned talent Whether you’re just starting your open source journey or looking to expand, this panel will offer actionable insights from organizations at the forefront of financial open source transformation. #FINOS #OpenSource #Finance #OSFF

39

1 Comment

There's a renewed focus on developer productivity in many organisations, especially with AI coding assistants promising massive gains. But it only really matters if we can measure developer productivity properly in the first place, and then notice how AI is actually shifting things. Laura Tacho, CTO at DX, and someone whose work I am a huge fan of, has an online course on measuring developer productivity and AI impact. But as often happens, the timing wasn't great for those of us in the southern hemisphere. Thankfully, Laura has just added a session that does work for Australia/NZ! 📅 Thursday Sept 4th 🕔 17:00 Sydney / 19:00 Auckland It's a 1-hour live coaching session plus 15+ self-paced videos, a workbook, and a reading list. Link in the comments.

22

7 Comments

Based on current market trends and recent tech lay-offs, it’s increasingly probable that we’ll see: ❌ Losers (likely to struggle): Generic full-stack developers Junior front-end/back-end roles Traditional DevOps engineers Manual QA testers Project managers without technical depth …especially with hiring freezes disguised as “strategic plans,” full-time roles converted to contract, and offshoring replacing entire U.S. teams. ✅ Winners (better prospects): AI specialists shipping production-ready models Cloud-security experts Cost-optimization engineers saving millions Data engineers building AI infrastructure Technical leaders automating entire teams 💡 As a technology leader, your mission is to transform your team so that more of them are in the “winners” category.

49

7 Comments

Managing Infrastructure as Code (IaC) often feels like a trade-off between speed, governance, and tool compatibility! If you’re exploring smarter ways to manage IaC and modernize CI/CD workflows, StackGen is worth a closer look. - and an AI powered! #IaC #DevOps #CICD #CloudEngineering #Automation #OpenSource #Innovation

22

1 Comment