Podcast
AI Upstarts
Hosted by Andrew Browning · EN
AI Upstarts is a podcast about the engineers who are actually using AI today, not the bosses talking about it in theory. Most conversations about AI in engineering happen at the top, in general terms. This one doesn't. AI Upstarts: Engineering's AI Innovators in Their Own Words goes straight to the people doing the real work: the engineers and environmental professionals quietly using AI to change how the job gets done, often well ahead of their own firms.
In each episode, one of them walks us through what they're using AI for, the tools and setup behind it, the difference it's made, and where they think it's all heading. No jargon, no sales pitch, just practitioners showing what they've built, in their own words. Hosted by Andrew Browning and James Griffith P.E.
2episodes
Episodes
Newest firstAll episodes
From Chat to Skills: James McWilliams P.E., Senior Traffic Engineer, TPD
3w ago00:16:20Tap to summarize"If you're just using the chat window, you're leaving 90% of AI's capabilities by the wayside."James McWilliams is exactly the kind of guest this show exists for: a licensed traffic engineer at TPD in Long Island, New York, who isn't theorizing about AI, he's building with it every day. Self-taught coder, systems thinker, and one of the sharpest practical minds we've had on. He shows what the other 90% actually looks like: context engineering over prompt engineering, and getting Claude to audit traffic signal plans for the quantity errors nobody catches by hand. Plus, his take on reusable skills, a quickfire round, and a genuinely fresh view of where engineering goes next.
Transcribe →"You've Got to AI Yourself": James Griffith P.E., Co-Founder & Chief Engineer, SchemeFlow
Jun 1500:10:18Tap to summarizeAI Upstarts goes straight to the engineers actually using AI today, in their own words. First up, James Griffith P.E., Co-Founder and Chief Engineer of SchemeFlow.James walks Andrew through how he actually uses Claude day to day. Drop an old report in, ask it to strip out the project-specific detail and generalize it for any project, and a few iterations later you have a template the whole team can reuse. It is the kind of systematic reuse engineering has talked about for years but rarely managed.He is just as candid about what the tech world gets wrong about the industry: that a building is not algorithmic, that two lines on a drawing can carry a hundred negotiations behind them, and the messy reality of getting things built.No theory, no pitch, just one engineer showing what he has built and where he thinks it is all heading.In this episode:From finishing Gaudí's Sagrada Família to engineering glass that survives earthquakesTurning a finished report into a project-agnostic template in minutes with ClaudePlugging AI into live data sources to auto-draft existing-conditions sections"First to be second": the innate caution of even the most innovative firmsThe one task James wishes AI would take off his plate for good
Transcribe →