Skip to Content
Event AgileDC Conference starts on Oct 26, 2026, 8:30:00 AM (America/New_York)
Practical Agility for Teams Using AI: Seven Years of Lessons from a Credit Union's ML Team
Location: Auditorium - 10/26/26, 11:30 AM - 10/26/26, 12:15 PM (America/New_York) (45 minutes)
Practical Agility for Teams Using AI: Seven Years of Lessons from a Credit Union's ML Team
Jonathan Bruskin and Richard Woolston
5719031340
Jonathan Bruskin and Richard Woolston
5719031340

Jonathan Bruskin is CEO and Co-Founder of Sprightbulb LLC, an Agile consulting and training firm that helps commercial and government organizations get their business and technology teams speaking the same language. Since 2001, he has worked with clients on product management, organizational transformation, and Agile coaching — building roadmaps, launching and rebooting teams, and turning strategy into outcomes that actually move the needle. Through its Sprightbulb Learning arm, the firm also delivers Agile training and certification, including certified courses through Scrum Alliance and Kanban University. Jonathan holds an M.A. in Communication, Culture, and Technology from Georgetown University (1999). Outside of work, he is the drummer for Siamese Dreamers, a Smashing Pumpkins tribute band.

Richard Woolston is an AI and Machine Learning leader with extensive experience designing and deploying enterprise-scale data and artificial intelligence solutions. He currently serves at America First Credit Union (AFCU), where he leads initiatives focused on building production AI platforms, machine learning infrastructure, MLOps, and data engineering capabilities that deliver secure, scalable, and measurable business value. Richard combines industry experience with a passion for education, teaching courses in machine learning operations while helping students bridge the gap between theory and real-world application.


America First Credit Union has been using machine learning (ML) in a production capacity for seven years - well before generative AI showed up. The ML operations team is surrounded by structurally waterfall organization whose broader Agile push has struggled to stick. Along the way, Richard Woolston's team discovered something the audience will recognize: traditional Agile practices were built for software engineering, and ML work is structurally different. The work is invisible when it's exploratory. The team is heterogeneous and running different cadences internally. Business stakeholders don't know what to ask for because they've never had the capability before. There's no SDLC for experimental work. And you cannot train a model without defining an outcome, which means outcome-thinking is a precondition.

In this session, Richard walks through five inflection points in the team's seven-year journey: the founding contracts, the hybrid Scrum-Kanban evolution, the "authorized refusal" behavior that protects against shipping the wrong model, the organizational friction of being and running Agile inside a waterfall org, and the OKR discipline he wishes he'd pushed harder sooner. Jonathan Bruskin frames each story and closes with usable takeaways.