Isaiah A.’s Post

Over the past two weeks, I've had the privilege of speaking at two key events about value-driven approaches in Data & AI. Here are some insights that resonated across both forums. 🌟 At the London Enterprise Tech Meetup: Hosted by Ian Ellis at the prestigious London Stock Exchange, this meetup brought together over 100 IT leaders. While the focus was on FinOps, I highlighted the crucial link between FinOps and a broad portfolio of Digital, Data & AI use cases and outcomes. The key takeaway? Understanding the business value is essential to assess spending coherence and identify cost-saving opportunities. 💡 Key Insights: - The importance of integrating FinOps with overall business strategy. 💼 - Assessing cloud investments for maximum value and cost optimization. ☁️💰 - FinOps as a tool for ROI evaluation and guiding further optimizations. 📈 🗼 At the datacraft event in Paris: This gathering of Data Governance Leaders, featuring insights from Xavier Maroteaux, focused on building an efficient data governance practice aligned with business value. My contribution was to share frameworks and guidelines for aligning data governance with high-value use cases. 💡 Key Insights: - Building awareness and an operating model for data governance. 🛠️ - Prioritizing and aligning data governance with business strategy for maximum impact... and making the case for adequate funding. 🔍 🔗 Connecting the Dots: The main commonality between these sessions? The imperative of a value-driven approach. By starting with expected outcomes, we can align operating models, practices, and value streams effectively. With this framework, we can clearly define requirements for both FinOps and data governance – essential components in delivering digital, data & AI products. What are your thoughts on integrating these approaches in your organization? #DataGovernance #FinOps #ValueDrivenStrategy #DigitalTransformation (Thanks for the invitations & organization Ian Ellis, Xavier Lioneton, Isabelle Hilali, Caroline Daudeteau, Rob Howes)

The #1 FinOps challenge is "getting engineers to take action on cost optimization" according to the FinOps Foundation’s State of FinOps 2023 survey. It's not that they're averse to saving 💰. They simply don’t have the insights needed to optimize costs. Cost optimization should be easy. 🍃 And it can be—with AI. Dive into the heart of the "#1 FinOps Challenge" now: https://lnkd.in/gBn2k2y5 #FinOps #CostOptimization #DataObservability #DataFinOps #dataobservability #cloudoptimization #observabilitytool #observabilityplatform #analytics #aiplatform #unravelplatform #dataplatform #cloudops #optimization

@Anodot’s Roundtable Recap: Leading Experts Discuss Cost Control, AI-driven FinOps, and the Future of FinOps Sean Donaldson Inbal Tadeski https://lnkd.in/d9_beJi4

AI is expensive, and it's coming to consume your data budgets. Farial Shahnaz and I recently published a blog on how you can (and should!) apply FinOps principles to your data spend too. Looking forward to chatting about this, and more, at the FinOps Foundation's FinOpsX conference this week!

finops tools today all suffer from the same problem. it is HARD to separate the signal from the noise. 🚫 you get dozens of recommendations a day. 😵💫 or worse, have to sift through foreign visuals on a dashboard (some even change the naming convention of the pricing data!). then you have to understand which of those recommendations it makes sense to take action on... THEN you have to actually do the work to realize the value. so we got to thinking... what if you could combine all the steps in the finops process? in other words, what if you could make it TRIVIAL to: 🔵 detect misconfigurations 🔵 extract precise text AND numeric-based insights relevant to YOU 🔵 then automatically generate code to solve YOUR specific use case 🚨 now you can, with BlueArch's InfraGPT. 🚨 Dieter thinks it's pretty cool (obviously so do i, but i'm biased 😊) and i would be remiss if i didn't offer a special thanks to the team that made this happen: David, Maria, Isaiah, Mayank and all of our early customers (Bjorn, Anand), investors (Kris, Rick) and advisors (Sairam, Ryan); it takes a village. the way SAs design and build cloud infra has just fundamentally changed. but dont take my word for it, reach out to someone on team to book a demo to see for yourself. LFG! 💙

Exploring MLOps: Where Data and Deployment Converge. Just finished MLOps concept from DataCamp, and I'm thrilled to dive deeper into the world where data-driven strategies meet seamless deployment.

“If business analytics drove the first wave of the data infrastructure evolution, operational analytics will drive the next wave, with Apache Flink at the center of it. Enterprises will begin en masse adopting real-time and streaming capabilities to begin using data to drive operational decision making. This will create new value in areas such as logistics, financial transactions, inventory management and fraud detection”.

Recently Haytham Abuelfutuh presented about Batch Workloads and how Union and Flyte are used for Large scale batch and near realtime workloads. Check out this post to 1. Learn about the challenges in using Kubernetes for running Batch workloads 2. The problems in the programming model with Batch workloads. Often times its not a single job - but it may start with one 3. The simplification that can be introduced by using Union. Union runs them across multiple clouds, multiple kubernetes clusters, securely accessing large amounts of data, while maintaining efficiency through - caching, smart techniques in which container state is saved, how pods are used, how back pressure is maintained etc. Check out the blog.

Venturing into the depths of Kubernetes' etcd, my curiosity was piqued by the enigmatic "Raft" algorithm. Join me on this journey as we unreval how Raft algo works in the realm of orchestration! 🚀 #Kubernetes #etcd #RaftAlgorithm Leader Election: Imagine a group choosing a leader. Raft does just that. In the etcd cluster, a "leader" node emerges to guide write operations and ensure data consistency. Think of it as the captain steering the ship! Log Replication: When a write request occurs, the leader logs it and shares with others. Once most agree, the entry becomes "committed," and data stays in sync. Handling Failures: Leaders can have coffee breaks, or even nod off. Raft handles this smoothly. If a leader goes offline, a new one is elected. The node with the most up-to-date log steps up, ensuring the show goes on without a hitch. Term-based Safety: Raft operates in "terms," marking time periods. Each term has one leader, ensuring order. If a higher-term leader appears, others gracefully step aside. It's like an organized relay race for leadership! Data Resilience and Reliability: The climax of our journey! With Raft, etcd promises data resilience and reliability. The synchronized dance of log replication, leader election, and commitment creates a solid base for Kubernetes' operations. Unlocking the Raft algorithm's role in Kubernetes' orchestration is like finding the key to a puzzle. It's the unsung hero that ensures all the pieces fall into place, harmonizing the data ensemble. Let's keep exploring 😀

For those curious, I wrote up a simplified overview of the considerations for building machine learning infrastructure and how it differs from building other software platforms. Check it out: https://lnkd.in/ge_4ZDNB