Anthill Inside 2019

A conference on AI and Deep Learning

Tickets Propose a session

Recommendation @ Scale

Submitted by Aditya Patel (@adityap) on Tuesday, 30 April 2019

Section: Crisp talk Technical level: Intermediate Session type: Lecture

Abstract

Recommendation is one of the most traditional and wide spread use case of Machine Learning. In this talk we want to showcase, how an advanced recommendation engine can be served at scale in Glance. Glance is an AI-powered, content driven, personalised Screen Zero (lockscreen) platform for mobile, which is used by over 26M DAU users in India. The talk will take you through each component of a recommendation engine and in the end will showcase the learnings which we got from our experiments to make it functional at scale.

Outline

  • What is Glance
  • Recommendation Engine in Glance
  • Serving Architecture
  • Learnings

Speaker bio

Aditya Patel is Director, Data Science at InMobi-Glance. Previously he was head of data science at Stasis and has 7+ years of experience spanning over the fields of Machine Learning and Signal Processing. He graduated with Dual Master’s degree in Biomedical and Electrical Engineering from the University of Southern California. He has presented his work in Machine learning at multiple peer reviewed conference. He also contributed to first generation “Artificial Pancreas” project in Medtronic, Los Angeles. In his current role, he is aiding in building the biggest content platform in India.

Links

Slides

https://www.slideshare.net/secret/rvHf6YvkfSKvFx

Comments

  • Abhishek Balaji (@booleanbalaji) Reviewer 5 months ago

    Hi Aditya,

    Thank you for submitting a proposal. The proposal in its current form is too Glance-specific and does not elaborate on a larger problem statement. Here’s some feedback + next steps for your proposal:

    1. The proposed talk cannot be only about Glance. Talk about some other application of building such a recommendation engine.
    2. It seems that Glance is mobile only, but the proposal is missing the unique challenges which come with ML models on mobile. Do elaborate more on the problem of vernacular languages and the considerations when designing for mobile.
    3. What was the context/problem/requirement that led Inmobi to build this recommendation engine or Glance.
    4. Why were solutions available in the market not customizable or adaptable for InMobi’s use case? What metrics did you use to evaluate these solutions before deciding to build a custom solution?
    5. Why did you decide on build versus buy? What approaches did you consider before building out Glance the way you did? Show detailed comparisons.
    6. Deep dive into the solution you have built, including show the various stages of iteration and the challenges you have faced at different stages of Sigma’s evolution. e. Show before-and-after situation comparison – what was life in InMobi before Glance and after Glance? What was your innovation’s big win?

    Next steps: The current proposal is too specific too one tool and lacks details for us to make a decision on your talk. Add more details to your slides based on the feedback above and update them here before 21 May, so we can close the decision on your proposal.

  • Aditya Patel (@adityap) Proposer Reviewer 5 months ago

    Hi Abhishek,
    Updated the slides with suggested changes. Let me know if you have more questions.
    Thank,
    Aditya

Login with Twitter or Google to leave a comment