Loblaws is one of Canada’s largest grocery store chains, Mefta's team at Loblaw Digital runs several ML systems such as search, recommendations, inventory, and labor prediction on production. In this conversation, he shares his experience setting up their ML platform on GCP using Vertex AI and open-source tools. 

The goal of this platform is to help all the data science teams within their organization to take ML projects from EDA to production rapidly while ensuring end-to-end tracking of these ML pipelines. We also talk about our overall platform architecture and how the MLOps tools fit into the end-to-end ML pipeline.

//Bio
Mefta Sadat is a Senior ML Engineer at Loblaw Digital. He has been here for over three years building the Data Engineering and Machine Learning platform. He focuses on productionizing ML services, tools, and data pipelines. Previously Mefta worked at a Toronto-based Video Streaming Company and designed and built the recommendation system for the Zoneify App from scratch. He received his MSc in Computer Science from Ryerson University focusing on research to mitigate risk in Software Engineering using ML.

--------------- ✌️Connect With Us ✌️ -------------
Join our slack community: https://go.mlops.community/slack
Follow us on Twitter: @mlopscommunity
Sign up for the next meetup: https://go.mlops.community/register
Catch all episodes, blogs, newsletter and more: https://mlops.community/

Timestamps:
[00:00] Introduction to Mefta Sadat
[01:04] Mefta's background  
[02:45] Mefta's journey in ML Engineering
[04:19] Use cases of Machine Learning at Loblaws
[06:00] Loblaws' team operation
[07:37] Number of people in the team and number of users in the platform
[08:40] Software engineering process
[10:47] Data platform vs ML platform
[13:10] Timeline leveraging machine learning in Loblaws products and business
[15:01] Transition from legacy systems to the cloud
[16:47] Recommendation System use case - Legacy Style Stack and its impact on the business
[21:01] Biggest challenges and pain points
[24:31] Choices of tools to use  
[27:31] Dealing with data access
[30:39] The good, the bad, and the ugly  
[32:48] Setting up alerts on image classification models
[33:53] Productionizing ML passion
[36:00] Post-deployment monitoring of recommendation systems
[37:47] Wrap up