Pittsburgh TechFest 2026

I run Lead Singer Wanted, a Pittsburgh-based live band karaoke act: the audience picks the songs, gets on stage, and sings lead while a real band plays behind them. I also wrote the software that runs the show, including QR-code signups, the singer queue, synchronized lyrics, the sheet music on the musicians' iPads, and the video wall. Oh yeah, and a Guitar Hero controller drives the whole thing!


The problem: five screens, four musicians, and a crowd all have to agree on what is happening right now. Every part of that can fail mid-song, and there is no maintenance window. When something breaks, the next song is only seconds away.


The solution turned out to be old-fashioned engineering, not fancy tools: one small event driving every device, a deliberately dumb server, recovery plans written before the failure happens, and a human kept firmly in charge of anything involving taste, because a perfectly "fair" queue can still put five slow songs in a row and empty the dance floor.


The impact: a side project that became a production system, 150+ shows over seven years, and engineering habits I now use wherever software meets real people.


This is a live demo, not just a slide deck. You'll scan the QR code from your seat, sign up like a real audience member, and we'll build a working queue in the room. And then I'll show you what happens when it breaks! You'll leave with engineering habits you can use immediately. And probably a song stuck in your head.


You don't need permission, a big team, or a maintenance window to build real production software. Pick something you love and let real users break it. The show must go on!

As extreme weather, wildfire risks, aging infrastructure, and growing electricity demand place new pressure on the power grid, utilities and communities need better tools for resilient decision-making. In this talk, I will discuss how human-AI collaboration can support grid infrastructure planning and operations by combining data-driven prediction, uncertainty quantification, and optimization with human expertise. Drawing on recent research in outage analytics, distributed energy adoption, wildfire risk, and grid resilience, I will show how AI can help decision-makers anticipate infrastructure vulnerabilities, evaluate tradeoffs, and respond more effectively under uncertainty. The talk will emphasize a practical question: how do we design AI systems that do not replace human judgment, but instead make expert decisions more reliable, transparent, and scalable?

Someone at your company got handed AI and told to "go figure it out." More and more, that someone is a technologist — and the instinct is to prove it by mastering the tools. That's the fast climb up "Mount Stupid," where quiet dabbling turns into shadow AI, exposed data, and surprise costs nobody budgeted for.
This talk reframes the job. The strongest AI champions aren't the most technical people in the room; they're the ones who apply judgment to real business problems. We'll break down what the role actually demands: justifying the investment to leadership, disarming pushback from other teams, finding where the value truly is, and catching the quiet killers — data, process, and legal readiness — before they sink an initiative.
You'll leave able to drive AI adoption that delivers instead of detonates, and to become the person who gets the credit for bringing it in right.