From POC to Multi-Million Dollar Investment: Rapid Scaling of AI Prototypes
A quick AI prototype is the best way to get buy-in for your great new concept! But after you have that green light from your enterprise, how do you scale? T-Mobile’s enterprise AI team began as a small, 12 week prototype – and now is a fully enterprise-essential team with multiple projects spanning business units and dozens of engineers. In this talk, Heather will guide us through the lessons she’s learned in her 4 years of experience designing prototypes, running POCs, and scaling successful experiments to create software that makes millions of real-time predictions a day.
Key Takeaways: • In agile POCs, people tend not to write logs – but they’re essential to iterating on your prototype at all. Logging must be MVP for AI prototypes.
• Design must be agile and run alongside dev teams. Agile design must be informed by data.
• Make your teams fully stacked – from analyst to DevOps engineer – to avoid wasting cycles coordinating with other teams for essential functions.
Heather Nolis is a founding member of the AI @ T-Mobile team, focusing the conversion of cutting-edge analyses to real-time, scalable data-driven products. She spends her time irl with her number-obsessed son Amber in rainy Seattle. You can find her @heatherklus on Twitter, where she talks about super sweet machine learning, data for good, and bad reality television.
