There was a fantastic crowd for the Bots + AI August 28th “Driving eCommerce Sales with AI” event at the Betaworks Studios offices. In addition to our usual eager to learn developers and product in the crowd, the visual search and fashion topic brought out quite a few fashion technology entrepreneurs including a prize winner from an IBM contest.
Luke Schantz from IBM led off the presentations of the night sharing a really sharp visualization of momentum shifts from the US Open SlamTracker. Behind the scenes is a data pipeline from the referee’s office and sensors through IBM services all the way to push notifications and highlights on fan phones.
The next IBM Data Science event August 30th promises to share more of the internals and will also have a few lucky attendees winning US Open tickets.
Joseph Babcock, Lead — Data & Experimentation at Jetblack (a Walmart incubator company) was the next speaker and shared lessons learned from his former role at Netflix as well as Jetblack. Getting further context on why a user made a decision through a conversational medium can fill in the predictive gaps and allows JetBlack to personalize even further. Instead of just guessing why a decision was made they can ask. It’s a double edged sword collecting more user preference information as the user expectations rise for you to be better at matching their taste!
Finally, Balaji Ravindran, COO of Markable.ai presented on computer vision and deep learning AI for fashion retailers. Text search for a particular style doesn’t lend itself to fashion and often gives zero or disappointing results. Visual search can provide immediate and targeted recommendations. Manually curated, pre-built fashion collections don’t account for out-of-stock items or a dynamic hybrid of users selecting some items but wanting ideas for others like a live human can. An AI can scale that conversation with millions of people.
After the presentations there was a quick audience asks segment where over a dozen audience members participated including a few studying machine learning and looking for internships, talented developers looking for a new opportunity, and startups and others looking for new members. The conversations carried on until closing time and a few ventured further outside to continue.
