Microsoft Coaches Drive Momentum at Volaris Revenue Signals Accelerator

Another CSI Volaris Group accelerator in the books this time in West Palm Beach. After participating in several of these, a few lessons consistently stand out: 1. Network, network, network. Some of the most valuable outcomes come from conversations outside the formal sessions. Connecting with GMs, portfolio leaders, and transformation teams often uncovers opportunities and ideas that wouldn't surface otherwise. 2. Don't explain agentic develoment - demonstrate it. For teams new to agentic DevOps, live demos and co-building sessions create understanding and confidence much faster than slide decks. 3. Optionality matters. During the accelerator, one of the AI services (A\) some were using became temporarily unavailable. Teams with access to multiple models and providers were able to continue working without interruption. AI resilience is becoming a strategic capability. This is why many teams like to use a multi-model agentic harness like GitHub Copilot 4. Take care of yourself. These are intense, dawn-to-dusk days. I've found that even 30–45 minutes of exercise and proper sleep significantly improves the energy and focus I bring to customers and teammates. Big thanks to the Microsoft coaches and teams who help make these accelerators successful. Balancing this work alongside ongoing customer commitments isn't easy, but seeing organizations accelerate their developer transformation makes it incredibly rewarding. MS Team: Hammad Aslam, Cody Carlson, Krystal Folkes Brown, Kaan Turgut, Dan Partington, CPA, CMA, Anurag Mathur, Jimmy Shah

Just back from developer heaven—aka GitHub Universe! The biggest takeaway? Orchestration. The future of code isn't written, it's orchestrated. GHU made it clear that devs are shifting from coders to conductors of AI. The #1 contributor at Github engineering is now GitHub Copilot! 😲 Here’s are my top announcements that make me bullish on Github: - Agent HQ - making agents native to the Github flow - can govern agent workflows at scale - Customizable agentic code review and issueOps agents (automate your boring work and steer your agent workforce for you) - Plan Mode in VS Code copilot - OpenAI Codex is already available in VS Code Insiders now! (why have one - agent when you can have 4 in parallel as you code?) - Better copilot metrics dashboard and API (finally) Also, here are wild Github stats I’ve discovered: - 1 dev/sec joins Github - 230 repos/minute are created - #1 internal contributor at Github Eng is GH Copilot - #1 language on Github is Typescript (thanks to the boon in AI) It was a great experience to be at the center of these announcements and connect with pretty smart technical founders and hackers. A huge thank you to the GitHub for Startups team (Garrett Wood Corey Dolik Ariana Song) for the opportunity—it was a privilege to meet the team and the other founders + devs in the GitHub community. ⭐ Looking forward to championing Github for devs across the spectrum! #GitHubUniverse #AI #GitHubCopilot #agenticdevops

I Built Context-Aware CI with GitHub Copilot SDK and Microsoft WorkIQ - it's the ability to catch code issues that violate what was said in a meeting, email, etc. We've all seen it: 2-hour planning meeting, team agrees on PostgreSQL, then someone opens a PR adding MongoDB two weeks later. 40 hours of rework. The root cause? Context fragmentation. Decisions live in Teams transcripts, emails, and "someone's head"—but CI only knows code. What Microsoft + GitHub makes possible: An integrated, end-to-end business layer for development context. WorkIQ surfaces your M365 data (meetings, emails, Teams, documents). The GitHub Copilot SDK provides the reasoning & agentic engine. Together, they close the loop between organizational decisions and code automatically. I built a proof-of-concept that changes this: GitHub Copilot SDK + Microsoft WorkIQ = CI pipeline that validates code PRs against your team's actual decisions. 🔹 Developer wants to merge code via PR 🔹 CI extracts context, queries relevant meeting transcripts 🔹 Copilot SDK analyzes diff against decisions 🔹 Three-state outcome: ✅ PASS / ⚠️ WARN / ❌ FAIL 🔹 PR comment outline the violations to specific meetings as an example. The result? Rework incidents drop. Reviews focus on logic, not policy. Institutional knowledge becomes enforceable not just documented. The future of agentic devOps & CI isn't just testing code — it's validating intent.  And it's only possible when your business context and the dev toolchain are truly integrated. My full technical writeup with implementation details is in the thread below 👇

163 Security Vulnerabilities. That’s where "Vibe Coding" got us. Here is how we fixed it with Vibe Engineering. TL;DR for the busy person - 🛑 Vibe Coding: Prompting based on intuition (Single Player Mode). Fast, but dangerous. E.g: Prompt, Paste, Pray. - ✅ Vibe Engineering: Orchestrating agents with repo context (Multiplayer Mode). Secure and scalable. Had an incredible time this past week speaking at Vancouver.dev & Vancouver #Azure User Group! 🎤 We talked about the elephant in the room: AI helps us code faster, but does it help us code safer? At my last company, we "vibe coded" our way to an MVP. It felt like magic until the security audit came back with 163 vulnerabilities. The AI wasn't bad - it just had Context Amnesia & and we didn't have mature AI SDLC practices. It didn't know our security rules, so it broke them. The shift to Vibe Engineering: Instead of just prompting a chatbot, we started prompting the Repo. By using Context Engineering primitives (like .github/copilot-instructions.md, *.agents.md), we moved from a "Typist/Prompter" mindset to an "Orchestrator" mindset. The result (as seen on screen 📸): We assigned a security vulnerability to a custom @security-agent. It analyzed the code, proposed a fix, ran the tests, and opened the PR you see behind me. Zero human code written, 100% human architectural review. Vibe coding is fun for hackathons. Vibe Engineering is for production. Thanks to the organizers for having me too! 🚀 #AI #GitHubCopilot #DevOps #VibeEngineering #Microsoft

Made it to 6 months at Microsoft! 🕺 Here are my top few lessons so far: - Scale through others and through systems. Great partnerships and reusable assets go further than individual effort. - Intrapreneurship is a huge unlock. Build, test, get feedback, repeat. - Mindset matters. Optimism and belief in what’s possible genuinely change outcomes. - Optimism is underrated. The right mindset changes how you solve problems. - Enablement beats just selling tools. Adoption happens when you help people build capability, not just turn something on. - Governance is an accelerator when done right. The goal isn’t slower delivery - it’s trusted delivery at scale. - Enterprise work can be deeply rewarding. It takes longer, but the relationships and long-term impact are the highlights for sure. Still learning. Still grateful. Still building. And yes, that's my little awesome niece 💛 next to me, and I gave her a tour of the Microsoft HQ at Redmond last week! :)

Yesterday was a special day as I had the privilege of spending time with Jeffrey Snover, discussing how professionals should navigate their careers in the age of AI, among other topics. Here are a few key ideas that resonated with me: 🔹 Write career plans in pencil. AI is highly disruptive, and adaptability matters more than rigid long-term plans. 🔹 Expect a three-year period of turmoil. Markets and roles will shift faster than new ones are created. 🔹 Engineering will increasingly become a commodity. As pure engineering becomes cheaper, value moves elsewhere. 🔹 Solve for the customer. Professional value is created by deeply understanding customer needs and applying AI to drive outcomes. 🔹 Thought is the driver of prosperity. As AI lowers execution costs, original thinking and judgment become the true differentiators. 🔹 Ambition has a cost. High-level success demands trade-offs, and being intentional about those choices matters. The main takeaway for me is that in an AI-first world, value shifts from “what you build” to “how you think.” I am grateful for the conversation and the perspective it provided. There is much to reflect on. Thank you again Jeffrey Snover for your inspiration and valuable contributions, which have greatly influenced my journey and will continue to guide me in the future. Your remarkable impact has supported my growth from being a technical author to an early adopter of various Microsoft technologies. Looking forward to inviting you to Toronto for an event.

My super TL;DR on my VanJS talk! How We 2x’d Dev Velocity (and supercharged our AI copilot) Our MVP was sinking under a “perfect” architecture—CQRS/Event Bus, max standards, minimum speed. This was an awesome choice for a different product we had then it changed. Frequent pivots made every change a headache and slowed us. We needed speed + safety, so we spiked three options: - tRPC: Great type safety, but too coupled. - Traditional REST: Flexible, but brittle and bug-prone. - ts-rest: ✅ The Goldilocks solution. Why ts-rest + Next.js won: - Contract-driven dev: One contract.ts = single source of truth for FE + BE. - End-to-end type safety without tight coupling. - DevEx boost: Full stack runs locally with pnpm run dev. Our Github Copilot was supercharged: Co-located contract + FE + BE code gave GitHub Copilot much better context → smarter code gen, debugging, and refactoring. How you structure your repos matters in the age of AI. Impact: 🚀 - 2x faster feature delivery - 🐛 ~30 fewer FE/BE bugs per week - ✅ MVP shipped on time No silver bullet for architecture and really depends on the requirements, team skill set etc. Takeaways for devs & tech leads: - Optimize for your current reality—not theoretical perfection. - Shared contracts = fewer bugs, happier teams. - Good architecture makes your AI copilot so much more effective. #vanjs #github #githubcopilot #tsrest #typescript

I completed 15 years at Microsoft. Five teams, seven managers and four roles later - here's my unfiltered advice for succeeding in corporate while staying sane. 1. 𝐘𝐨𝐮𝐫 𝐜𝐚𝐫𝐞𝐞𝐫 𝐢𝐬 𝐚𝐧 𝐚𝐥𝐛𝐮𝐦. Some songs will be hits, some will be inconsequential, some will be straight up bad. Learn to be okay with downturns and keep moving. 𝐴 𝑏𝑎𝑑 𝑟𝑒𝑣𝑖𝑒𝑤 𝑠ℎ𝑜𝑢𝑙𝑑 𝑛𝑜𝑡 𝑑𝑒𝑓𝑖𝑛𝑒 𝑦𝑜𝑢𝑟 𝑠𝑒𝑙𝑓-𝑤𝑜𝑟𝑡ℎ. 2. 𝐓𝐫𝐞𝐚𝐭 𝐲𝐨𝐮𝐫 𝐦𝐚𝐧𝐚𝐠𝐞𝐫 𝐚𝐬 𝐚 𝐜𝐚𝐫𝐞𝐞𝐫/𝐥𝐞𝐚𝐝𝐞𝐫𝐬𝐡𝐢𝐩 𝐜𝐨𝐚𝐜𝐡 𝐲𝐨𝐮'𝐫𝐞 𝐛𝐚𝐬𝐢𝐜𝐚𝐥𝐥𝐲 𝐠𝐞𝐭𝐭𝐢𝐧𝐠 𝐟𝐨𝐫 𝐟𝐫𝐞𝐞. Don't waste your precious time, working for losers. I’ve learned more by working for leaders I aspire to be like than from any course I’ve ever taken. A great manager will advocate for you when you're not in the room and will teach you how to operate when you eventually get into that room. 3. 𝐊𝐧𝐨𝐰𝐢𝐧𝐠 𝐰𝐡𝐞𝐧 𝐭𝐨 𝐪𝐮𝐢𝐭 𝐢𝐬 𝐚𝐧 𝐮𝐧𝐝𝐞𝐫𝐫𝐚𝐭𝐞𝐝 𝐬𝐮𝐜𝐜𝐞𝐬𝐬 𝐜𝐚𝐭𝐚𝐥𝐲𝐬𝐭. Watch for three things: is your scope actually growing, are you still learning, are you still having fun. If two out of three are false, it's your sign to change the role, team or company. 4. 𝐁𝐞𝐢𝐧𝐠 𝐭𝐡𝐞 𝐬𝐦𝐚𝐫𝐭𝐞𝐬𝐭 𝐨𝐧 𝐭𝐡𝐞 𝐭𝐞𝐚𝐦 𝐝𝐨𝐞𝐬𝐧'𝐭 𝐠𝐮𝐚𝐫𝐚𝐧𝐭𝐞𝐞 𝐟𝐚𝐬𝐭𝐞𝐫 𝐩𝐫𝐨𝐦𝐨𝐭𝐢𝐨𝐧 𝐯𝐞𝐥𝐨𝐜𝐢𝐭𝐲. I’ve seen people who are curious and have a growth mindset almost always outpace the most intellectually gifted people in terms of career growth. 5. Unless you're making pacemakers or rockets, 𝐧𝐨𝐭𝐡𝐢𝐧𝐠 𝐢𝐬 𝐚𝐬 𝐮𝐫𝐠𝐞𝐧𝐭 𝐚𝐬 𝐢𝐭 𝐬𝐞𝐞𝐦𝐬. Give priority to your family and health; the work will still be there when you log in tomorrow. 6. 𝐑𝐞𝐥𝐚𝐭𝐢𝐨𝐧𝐬𝐡𝐢𝐩𝐬 𝐦𝐚𝐭𝐭𝐞𝐫 𝐚𝐬 𝐦𝐮𝐜𝐡 𝐚𝐬 𝐭𝐡𝐞 𝐪𝐮𝐚𝐥𝐢𝐭𝐲 𝐨𝐟 𝐲𝐨𝐮𝐫 𝐰𝐨𝐫𝐤 𝐚𝐬 𝐲𝐨𝐮 𝐜𝐥𝐢𝐦𝐛 𝐭𝐡𝐞 𝐥𝐚𝐝𝐝𝐞𝐫. Everyone who reaches senior level is competent, that's a given. What separates the person who becomes a VP from the one who stagnates at Principal is the relationships they've built with the people who influence those decisions. Invest in networking, the right way. Fifteen years in, this is what I'd tell my younger self (picture 👇) who was over the moon to join her dream company fresh out of college. Gosh! What a joy that first-ever offer letter was.  Which of these resonates most with you? Let me know in the comments. #Microsoft #careerjourney #corporatelife

Three years ago, I left Microsoft. I thought my biggest challenge would be replacing my salary. I was wrong. The salary was the easy part. The hardest part was realizing how much of my success belonged to Microsoft, not to me. For 15 years, doors opened before I entered the room. People returned calls. Meetings got booked. Trust existed before the conversation started. I thought that was my ability to sell. Looking back, I was often selling from a position of borrowed credibility. The day I left, that credibility disappeared. Nobody knew my company, my work, or my name, and for the first time in my career, I had to build trust from zero. What I understand now is that companies do not just pay you a salary. They give you credibility, distribution, and opportunity, and inside a great company those things are so constant that you mistake them for your own. You only notice they were never yours when they are gone. Leaving Microsoft did not make me less capable. It removed the system that had been making my capability visible for fifteen years. I had spent that time building expertise without realizing that expertise and visibility are two completely different things, and only one of them leaves with you. Looking back, there are three things I wish I had started building much earlier: a personal brand that existed independently of my employer, sales skills built on earning trust rather than borrowing it, and a network outside the corporate world I had spent my career inside. Three years later, I have exceeded my corporate salary. But the lesson that mattered most was not financial. It was understanding that leverage takes years to build, and that a company can lend you credibility for as long as you work there, but only you can actually own it. If you're still inside a company, that ownership is worth starting now, while you still have the safety net underneath you.

🚀 Announcing a major milestone in Microsoft AI Today, Microsoft announced three new world‑class MAI models, now available in Microsoft Foundry, marking another big step toward delivering high‑performance, enterprise‑ready AI at scale. 🔹 MAI‑Transcribe‑1 – State‑of‑the‑art speech‑to‑text across 25 languages, built for real‑world audio with industry‑leading accuracy and speed. 🔹 MAI‑Voice‑1 – A top‑tier voice generation model that produces natural, expressive speech and enables secure custom voice creation with just seconds of audio. 🔹 MAI‑Image‑2 – A turbocharged image model delivering faster generation, high‑quality visuals, and clear in‑image text—already rolling out across Copilot, Bing, and PowerPoint. What stands out is the price‑to‑performance and responsible AI design—these models are optimized for efficiency, enterprise scale, and real‑world use cases, and are already powering products millions use every day. Excited to see how developers and enterprises build richer voice, image, and multimodal experiences using these capabilities in Foundry. 👉 Read more: https://lnkd.in/eynPNjFu #MicrosoftAI #AI #GenerativeAI #Foundry #Copilot #EnterpriseAI #Innovation