Payments Forecasting with Retail Data

A huge thank you to the committed folks who joined us at Code Talent on St. Patrick's Day for our Local Expert speaker series. Choosing to invest your holiday evening in a deep dive into applied ML made for a highly engaged and focused meetup.

We had the privilege of hosting the brilliant Bruce Hicks for his talk, "Payments Forecasting in the CPG and Retail sector." It was an amazing, highly math-focused discussion centered entirely around what it takes to do real AI at scale.

Bruce delivered a deeply technical, practitioner-focused deep dive into building insight-centric, drift-resistant revenue forecasting models for retail businesses. We explored the statistical foundations of revenue forecasting using macroeconomic indices and walked through the full modeling lifecycle. This included everything from addressing complex pitfalls like sampling bias and simultaneity, to constructing business-weighted indices, fitting regression models with proper diagnostics, and establishing advanced time series monitoring frameworks.

What made the evening truly special was how the conversation evolved. While the room—filled with data scientists, ML engineers, and analytics leaders—was highly engaged in the pursuit of precise (<1% annual error) forecasts, the discussion became about much more than just the math. We dug into the human element of engineering: how to effectively advocate for complex mathematics to less technical stakeholders, build trust in your models, and ultimately align these highly technical systems directly with overarching business goals.

Thank you to Bruce and to everyone who brought such high-level technical depth and strategic insight to the discussion. We look forward to seeing you at the next one!