AI in Production: A Machine Learning 3D Computer Vision System Deployed to AWS

A Machine Learning 3D Computer Vision System Deployed to AWS : A Data Point

Presenter: Aleksey Nozdryn-Plotnicki, Data Scientist and Product Manager, NGRAIN

Bio: Aleksey leads the creation of machine learning solutions. He engineers the models, faces the customer, coordinates the development team, and writes some production code, but does not describe himself as a Software Engineer. Perhaps a "real" Data Scientist, if such a thing exists.

Abstract: Aleksey shares some experience from his team's recent deployment of a machine learning computer vision system to the cloud (Amazon Web Services). Some aspects of the problem were unique, but broadly the approach should be informative, especially for anyone with the basic problem of having a model that needs to be "used".

The objective of the talk is to provide a reference point to better prepare you for when you sit down with your team to work out the requirements and design for your own implementation.

Some characteristics of the scenario: website front end; offline training; on-demand prediction; non-technical end user in real-world business process; large assets as input to prediction; computationally heavy feature engineering; 3D computer vision; relatively low volume; extensive custom code; open source prediction library. The talk will be unavoidably AWS-oriented, but in the time allotted we won't be getting too lost in the weeds of the detailed implementation.

Agenda:
17:45 - Doors Open & Mingle
18:30 - Doors Closed & Introductions
18:35 - Talk
19:25 - Questions & Discussion
19:45 - Close

Please RSVP responsibly and also note that the entrance to the building will be locked by 18:30 hrs.