Problem:
Myntra is one of the leading fashion e-commerce companies in India. Myntra is focused on delivering best-in-class customer experience for all the fashion lovers from browsing to purchase and post purchase experience. Myntra provides curated, customized shopping experience to every user by matching deep understanding of the user with deep expertise on fashion and trends.
Myntra is leveraging AI and ML to solve complex problems on various facets of e-commerce operations - store front, user shopping experience, supply chain (inbound and outbound), pricing etc. and use cases continue to expand. Myntra deployed and continues to deploy many advanced machine learning models to solve critical business problems in these areas.
Machine learning lifecycle - getting the right datasets, model training, deployment, serving (online and offline), integration with services, model health monitoring and alerting is extremely complex. This requires data science, data engineering, ML engineering, software engineering and MLOps practices to come together. There is no single or set of open source or 3rd party tools that provide the capabilities required to manage the end to end ML lifecycle for Myntra scale.

Solution:
Myntra Engineering team designed and built an in-house Machine Learning Platform, with the goal to support the entire machine learning lifecycle from model development to deployment, serving, model health, feedback loop and integration with applications. The machine learning platform is scalable to support online/offline / batch serving patterns and provides self serve, CICD and MLOPs capabilities.

Outline of the talk:
The Lifecycle of a Machine Learning model
Role of Machine Learning Platform at Myntra
Design, Architectural principles and Tech choices
Deep dive into the platform capabilities
Overview of an online serving use case deployed on the platform
Best practices and learnings
Looking ahead

This talk will be jointly presented by Narayana Pattipati(Senior Architect, Myntra Data & Machine Learning Platforms) and Meghamala Ulavapalle(Senior EM, Machine Learning Platform)