Back

Speaker "Michael Liu" Details Back

 

Topic

Building Real World AI Solutions

Abstract

Information is the new gold. Information extraction from various sources is an important factor in building successful business in many industries. However this is a notoriously difficult problem. At Ancestry we help our customers to build an accurate view of their family history. In this workshop we will review the challenges related to building an accurate information retrieval system from a very large collection of historical newspapers. In the first part of this workshop we will review the models related to article classification, image segmentation, and information retrieval. We will also build some of these models using deep learning. In the second part of the workshop we will discuss challenges related to the large scale cloud-based production deployment of the models we discuss in the part one. We will also build serving environment to process the models quickly and efficiently.
Who is this presentation for?
Data Scientists, ML developers, CS Students, AI/ML Leaders
Prerequisite knowledge:
None is required. To get the maximum value of this workshop recommended at least some of the following: basic understanding of DS development principles, deep learing, NLP, image processing models, familiarity with python and java programming
What you'll learn?
How to build end-to-end DS pipeline.

Profile

Michael is a Sr. Machine Learning Engineer at Ancestry.com. He has various experience in developing and deploying advanced ML solutions. Previously, he was at Scientific Computing and Imaging Institute (SCI Institute) and Lawrence Livermore National Laboratory (LLNL).