Conference Video: How AI is Solving the $160 Billion Food waste Problem: Lessons in Forecasting at Scale

This is Shawn Ling Ramirez’s presentation from WiDS Puget Sound Conference 2021. Enjoy!

Abstract:

Time series forecasting at scale can be done in many ways. I will talk about two common methods - boosted trees and Bayesian Hierarchical Time Series - their pros, cons, and some hard choices we've made to address machine learning bias, cold start, and COVID, as well as how to iterate at speed.

Bio:

Shawn is data scientist and tech strategy leader, the Head of Data Science at Shelf Engine, and a consultant for companies developing AI for Good data products. With over 15 years building teams and analyzing social behavior as a professor, she now plays in forecasting and optimization to solve the food waste problem.