June Andrews, AI Instruments
Stitch Fix
Abstract:
Working on the competitive edge of industry is a double edged sword.
Novel and intricate technical challenges arise continuously, but the pressure is to avoid and mitigate complications that develop from these challenges. This puts folks in industry in a position to glimpse what the next wave of innovations and research will look like for technology, but not the freedom to delve into the necessary research. In this talk, we will share challenges faced in industry that beg for additional research. We will cover Replication in Data Science or how asking the same question of different people results in wildly different answers.
The latest challenge is the integration of machine learning into human decision systems. Unfortunately, many integrated systems can fall into The Uncanny Valley of ML, where it would be better to have either people or ML making a decision, but not both. We will cover experiments highlighting the impact of not addressing these challenges, best practices to avoid the incorrect use of technology, and a wishlist for research on understanding data and its uses in the world.
Bio:
Dr. June Andrews works on AI Instruments at Stitch Fix. Previously, she led building a Monitoring & Diagnostics platform for GE’s airplane engines in use today. The platform has since been extended to turbines in renewable energy and power plants. At Pinterest, June created the Signals Program, a feature store, supporting over 50 ML engineers. At LinkedIn, she supported growth and engagement during a 50% gain in membership. June holds degrees in applied mathematics, computer science, and electrical engineering from UC Berkeley and Cornell. Her hobbies include teaching at Berkeley and serving on advisory boards.
Friday, November 15, 2019 at 11:00am to 12:00pm
Kelley Engineering Center, 1001
110 SW Park Terrace, Corvallis, OR 97331
Michael Slater
No recent activity