Shane Davis

This is a useful dashboard if you are considering what models to use for your agentic architecture implementations. Having the cost profile is handy as well. This is obviously no substitute for experimentation, but useful to have a *more* objective confirmation of what you are experiencing, or perhaps a useful guide on what to play with when getting started. Full link below.

34

2 Comments

Microsoft and the Minnesota Vikings built Viper, an Azure‑based platform that centralizes game and practice video, player‑tracking data, and historical records into a searchable, coach‑centric system that streamlines scouting, player evaluation, and game preparation to accelerate data‑driven decision‑making and competitive advantage. #MSFT #AI #Azure  https://lnkd.in/eVQDAXxv

6

I built a local‑first RAG chatbot with FastAPI + Chroma, and wired it up as an MCP server so any MCP‑enabled client (like Claude for VS Code) can search my knowledge base. 🚀 Highlights: - FastAPI endpoints: /chat (answers + citations), /ingest (text, file, or URL), /health, and a simple UI at /ui - Vector store: Chroma (persistent), with Sentence-Transformers (local) or OpenAI embeddings - Clear, inline citations, with a compact references section - PDF/Markdown/TXT ingestion using Unstructured; optional “inbox” folder watcher for auto‑ingest - Optional OpenAI for generation (with a graceful local stub fallback for dev) - Exposes MCP tools: add_document, search_knowledge, clear_corpus — great for IDE agents Code (⭐️ welcome): https://lnkd.in/gtsWjyaT #RAG #AI #MCP #FastAPI #VectorDB #Python #OpenAI #DeveloperTools

2

Java isn’t broken, but some issues have gone unresolved for decades. Java’s biggest pain points haven’t changed in over 20 years: null pointer exceptions and excessive boilerplate. That’s where JPlus steps in: a modern Java superset designed to boost productivity and safety without forcing anyone to abandon the JVM ecosystems they’ve spent decades investing in. JPlus tackles these head-on with:  • strict null-safety baked into the language  • automatic generation of repetitive boilerplate  • cleaner, more expressive constructs that stay 100% Java-compatible This isn’t a rewrite; it’s Java with guardrails and sanity. For engineering leaders balancing roadmap pressure, technical debt, and team productivity, that’s a compelling equation. https://lnkd.in/gAH-H8rY

5

1 Comment

Discover lessons learned from Kroger's EPM implementation—starting with foundational components and building upward—to drive value across the enterprise while keeping their teammates excited and wanting more: https://lnkd.in/gR8PYwNs

1

Enterprise Architecture from the CFO seat For a CFO, the value of Enterprise Architecture is not cost control. It’s cost visibility. Many of the most expensive decisions in an enterprise never appear as line items. They surface later as technical debt, redundant platforms, integration complexity, and rising run costs. Architecture adds value when it makes long-term cost visible at the moment decisions are made, not after capital is committed. It clarifies where savings are real versus deferred, where cost is shifted rather than reduced, and where short-term efficiency creates long-term obligation. It’s a form of financial risk management.

36

6 Comments

Too many IT leaders are stuck in defense mode. Blaming slow vendors. Blaming outdated org charts. Blaming legacy systems. Don’t play that game. Instead - architect systems that scale under pressure, build teams that lead without permission, and demand performance, not excuses, from every vendor.

38

I've been researching Context Engineering recently, and thinking about what does Context mean for Enterprise Architects. This has resulted in my latest insights, Enterprise Architecture’s Critical Role in Driving AI Value Summary : Organizations often fail to unlock real value from AI when they focus solely on models or agents. Heads of EA must provide the full enterprise context of people, processes, data and technology, along with adaptive cross-line-of-business governance to ensure AI initiatives succeed. To read more (subscription required) visit https://lnkd.in/edaNTwsp Not a Gartner client? To access free, compelling insights visit https://gtnr.it/GExpert

11

The "success to data" pivot has become a disturbing industry trend. Once a product captures the market, the company often shifts from service provider to data broker. This pivot forces customers to pay for API access for their own data or data that they helped create.

8

1 Comment

#agile- Process for process’ sake, or actually useful? I work across multiple projects for multiple clients, and here’s one truth I’ve learned: tasks make or break delivery. If they’re not broken down, allocated, or briefed correctly, things unravel fast. Yet most people treat tasks as “the work itself.” Before you know it, you’re creating a task to clean up your tasks. That’s neither fun nor productive. Here’s how I keep it useful (not just process for the sake of it): ✅ Keep tasks small and bite-sized ✅ Put larger pieces of work in their own category ✅ Allocate story points or time from past experience (if it’s brand new, treat it as a bigger category) ✅ Avoid overwhelm by working on one at a time, clear it off, or change priority Agile isn’t just about how we work or the boards that store the work, it’s behavioural. The power is in how we break down and act on work, not how pretty the process looks. What’s your take? Is Agile helping your team deliver better, or is it still just theatre. #worklife #agile #agilemsthodology #projects #projectmanagment

6

3 Comments

If you’re using the SQL Analytics Endpoint over Lakehouse data, file health and maintenance really matters. Christopher  Finlan's latest blog does a great job walking through how to ensure your data is in the right shape for fast SQL querying. Check it out https://lnkd.in/gqt4CE4x #MicrosoftFabric #Data #Analytics #SQL #DataEngineering

39

1 Comment

I’ve got it under control.” — Every CIO, right before IT strategy hits the floor. AI. Cloud. Cyber. Data. All in motion. But here’s the issue: 👉 We’re juggling technologies… not building coherence. In too many telcos: • AI has no production-grade data behind it • Cloud = lift & shift, not re-architecture • Data platforms are disconnected from operations • Tech debt keeps compounding silently Result? ➡️ More complexity ➡️ Slower delivery ➡️ No real differentiation The missing piece is not another initiative. It’s architectural control to deliver the expected strategic outcomes. You don’t become an AI-driven telco by adding more balls to juggle. 👉 You get there by re-architecting the system. #Telecom #CIO #OSS #BSS #AI #ODA #IntelliCom

3

Claude Code now has built-in LSP support. This will produce more consistent code changes, allowing for Spec Coding of larger projects. It is an important upgrade! Read about what that means here: https://lnkd.in/eThKmKAd

6

Your problem isn't the software. I learned how to code in Fortran on an IBM Mainframe in the basement of Loyola Academy back when I was in high school. I wrote in raw code, on punch cards. Developing a proper functional program took months and a lot of checking to see if that " ; " was in the right place with the right syntax - trust me, it never was. Today at nutrād, our software engineers leverage Databricks + Vibe Coding to create sophisticated, real-world solutions for complex data challenges in seconds. Omaxn's "Vibe Coding Unlocked" Tuesday night at the TechNexus Venture Collaborative highlighted the massive power of the Databricks Platform and why Vibe Coding is the future. Instead of humans speaking computer, the computer understands the human. And the platform now goes way beyond simple understanding - it suggests better ways to structure the code and solve the problem. So the enterprise problem today isn't software, it's the willingness to change manual processes and weave in better, software-based solutions. A willingness of the enterprise to embrace machine-driven solutions that perform tasks more reliably and more accurately. It's a harder challenge than I make it sound. But the challenge can be overcome; we can help. -------- Many thanks to Pete Wilkins for the Vibe Coding invite and our friends at Databricks (Ventures + Startup, Andrew Ferguson, Steve Sobel), AI experts Kevin Rasmussen, Sriharsha Tikkireddy. And, while I didn't see you at the event Michael J. Franklin, grateful for all you've done to connect us to the great folks at Databricks and within the Polsky Center at the University of Chicago.

9

5 Comments

Developer experience significantly impacts delivery speed, particularly at scale. {unscripted} is rolling into Chicago with an all-star panel featuring Travis Gosselin (SPS Commerce), Murray Butler (Thrivent), and Sreenath Gotur (Amazon Web Services (AWS)) to discuss how they’re rethinking internal tooling with platform engineering and AI at the center. From developer portals to API-first automation, they’ll share how treating developers like customers helps reduce friction, drive adoption, and unlock smarter workflows. Can't join us in Chicago? View the full list of cities and our virtual event. 👉 https://lnkd.in/gfp4SMfd

2

🚀 Stop celebrating output. Start measuring outcomes 🚀 Many teams still equate “busy” with “impact.” But real progress isn’t measured by how many features we ship — it’s measured by the value we create and the problems we solve. When we focus on outcomes, we build with purpose, prioritize what truly matters, and create meaningful change for our users and our business. 👉 Read full article here: (https://lnkd.in/eh5KabrU) Let’s build value, not only features

13

1 Comment

Remaming your children? Continually changing metadata terms in an enterprise creates significant, compound costs that go far beyond the initial effort of updating a spreadsheet or system, often leading to, for example, a 40% increase in data management costs. The financial and operational burdens include high manual labor costs (up to 16 hours/week per employee in some cases), degraded data trust, and severe regulatory risks.

5

🚨 Columbus builders and developers — T-2 days! Tuesday is the day. 🚨 Come see how Kiro, Amazon’s agentic AI-powered IDE, revolutionizes how you build and ship applications - from rapid vibe prototyping to structured, spec-driven development. In just 30 minutes, Girish Bhatia and I will walk through a live build, showing how quickly you can turn an idea into a working application. Curious what modern, AI-powered development actually looks like in practice? This session is for you. 👉 See you there. #AWS #GenAI #AIDevelopment #CloudComputing #ColumbusTech #AWSMeetup #Innovation #DeveloperTools #Kiro #FranklinUniversity

5