https://murateksi.com/Recent content on Murat EksiHugoen-usSun, 19 Apr 2026 00:00:00 +0000https://murateksi.com/colophon/Sun, 19 Apr 2026 00:00:00 +0000https://murateksi.com/colophon/This is a small custom Hugo theme, hand-built in one sitting and refined since. No framework, no build step, no JavaScript bundler. Typography Display and editorial type is Fraunces, set in its variable form with the optical-size axis tuned up at large sizes so the drawing stays sharp. Body and UI type is Geist. Monospace for dates, section numbers, and metadata is Geist Mono. All three are self-hosted as WOFF2 files, preloaded in the document head, and served with font-display: swap.https://murateksi.com/now/Sun, 19 Apr 2026 00:00:00 +0000https://murateksi.com/now/Updated April 2026. The pattern of this page follows nownownow.com. The quarter’s work The Nordics Professional Services practice is in the middle of a three-customer portfolio that closes in Q3. Two are GameTech, one is industrial. The commercial side is well-set. The architecture reviews are where the quarter gets won or lost. Shipping the practice playbook for how we sell platforms versus products, internally. Most of the deals I have seen close in the last twelve months closed because an architect said something specific and credible at the right meeting.https://murateksi.com/reading/Sun, 19 Apr 2026 00:00:00 +0000https://murateksi.com/reading/Three tiers, loosely. A handful of books that I have reread enough that they have earned a permanent shelf. Books I am in the middle of this quarter. Things I plan to pick up when the current pile thins. Perennial Claude Shannon, A Mathematical Theory of Communication. Still the clearest writing I have encountered on any hard topic. I reread it when I catch myself writing in jargon. Christopher Alexander, A Pattern Language.https://murateksi.com/posts/technical-seller-advantage/Wed, 08 Apr 2026 09:00:00 +0200https://murateksi.com/posts/technical-seller-advantage/There’s a persistent divide in enterprise technology between “the people who build” and “the people who sell.” Engineers look at salespeople with suspicion. Salespeople look at engineers as resources to be deployed on calls. Both sides lose. The professionals who create the most value, and build the most durable careers, are the ones who bridge this gap. Why technical depth changes the sales conversation A traditional salesperson can present capabilities. A technical seller can diagnose problems.https://murateksi.com/posts/building-ai-ready-architecture/Thu, 05 Mar 2026 09:00:00 +0100https://murateksi.com/posts/building-ai-ready-architecture/Every week I talk to engineering leaders who want to “add AI” to their platform. They’re thinking about models, frameworks, and which LLM provider to use. Those are all valid questions. But they’re the wrong first questions. The first question is: can your current architecture support AI workloads at all? The data layer problem AI models need data. Not just any data. They need consistent, well-structured, accessible data with clear lineage and governance.https://murateksi.com/posts/selling-cloud-transformation-to-executives/Thu, 22 Jan 2026 09:00:00 +0100https://murateksi.com/posts/selling-cloud-transformation-to-executives/Every enterprise executive has a cloud horror story. The migration that went 3x over budget. The lift-and-shift that lifted costs without shifting outcomes. The “cloud-native rewrite” that took two years and still isn’t done. When you walk into a room to propose a cloud transformation, you’re not starting from zero. You’re starting from negative trust. Here’s how to navigate that. Stop leading with technology Nobody in the C-suite cares about Kubernetes.https://murateksi.com/posts/enterprise-ai-adoption-reality/Thu, 18 Dec 2025 09:00:00 +0100https://murateksi.com/posts/enterprise-ai-adoption-reality/I’ve watched dozens of enterprise AI initiatives launch with excitement and stall within six months. The pattern is remarkably consistent. A compelling demo gets executive buy-in. A team gets funded. Then reality sets in. The model works on clean data. Production data isn’t clean. The data problem nobody wants to talk about Every AI conversation starts with “what model should we use?” The right first question is “what does our data actually look like?https://murateksi.com/posts/serverless-at-scale-lessons/Wed, 12 Nov 2025 09:00:00 +0100https://murateksi.com/posts/serverless-at-scale-lessons/Most serverless content focuses on the first deployment. The “Hello World” moment. But the interesting part starts when your platform handles 200K messages per second and your on-call engineer gets paged at 3 AM because a downstream service changed its payload format. Here’s what I’ve learned building serverless systems at enterprise scale. Cold starts are not your biggest problem Everyone asks about cold starts. Almost nobody asks about observability. In a distributed serverless architecture, your biggest operational challenge is understanding what happened when something goes wrong.https://murateksi.com/case-studies/nordic-telematics-serverless/Sun, 15 Sep 2024 09:00:00 +0200https://murateksi.com/case-studies/nordic-telematics-serverless/Context A large Nordic telematics provider ran a fleet-management platform for more than 1.6 million commercial vehicles. Each vehicle reports position, diagnostics, and driver behavior several times per minute. At peak, the ingestion pipeline handled about 200,000 messages per second. The existing stack, a mix of dedicated EC2 instances running Kafka and bespoke Java consumers, was seven years old. It worked. It also woke people up twice a week. The business wanted to add three new vehicle platforms in the next fiscal year, which would roughly triple the throughput.https://murateksi.com/posts/cloud-architecture-business-decisions/Wed, 10 Jul 2024 10:00:00 +0200https://murateksi.com/posts/cloud-architecture-business-decisions/There’s a persistent myth in technology organizations that architecture is a purely technical discipline, that decisions about service boundaries, data stores, and deployment strategies belong exclusively to engineers. In reality, every architectural choice is a business decision in disguise. The Cost of Abstraction When a team chooses between a managed service and a self-hosted alternative, they’re not just evaluating technical trade-offs. They’re making a statement about where the organization wants to invest its engineering hours.https://murateksi.com/case-studies/bundesliga-real-time-ml/Mon, 20 Nov 2023 10:00:00 +0100https://murateksi.com/case-studies/bundesliga-real-time-ml/Context The Deutsche Fußball Liga (DFL), which operates the Bundesliga and 2. Bundesliga, runs a digital products group called DFL Digital Sports. They license real-time match data and insights to broadcasters, clubs, and betting partners worldwide. They wanted to add machine-learned match signals, like expected goals, shot quality, pressing intensity, to the live data feed. The requirement was sub-two-second latency from event on the pitch to signal on the broadcast graphic.