• Machine Learning Engineer with 6+ years of experience in machine learning, deep learning and MLOPs.
    • Presently working in Gojek on building a Data Science Platform on top of Kubeflow to enable DS teams to take models to production faster.
    • Working on a real-time feature store to manage data pipelines and features required for running real-time ML models in production.
    • Prior to moving into building the platform worked as data scientist in user-growth team in Gopay ( financial arm of Gojek ) to build models for estimating Customer Lifetime Value to help business target and cater to user needs in a better way. We have been working on a model that has helped organisation champion organic growth and other growth efforts. This model has helped organisation bring down cost spent on promotions by 22% without taking hit on growth efforts.
    • Worked as data scientist in Gopay to build models for fraud and account takeover detection ( ATO ). This model predicts whether an incoming request is an ATO or not. This model has been running in real-time setting and has resulted in 33% reduction in tickets.
    • Before joining Gojek, worked in DeltaX and worked on Data Driven Attribution where we created data-driven models by borrowing ideas from Game Theory like Ordered Shapley Value to attribute conversions tracked by our system which served as building block for another model that takes into account converting and non-converting user journeys to calculate the importance of a campaign for the advertiser. It helped advertisers get better understanding of their audiences.
    • Also worked on budget distribution problem where we helped develop an algorithm that would distribute budgets among competing ad-groups running on Facebook while optimizing the business objective of the advertiser. This is a big step up from traditional approaches which mostly look at single indicator and it helped advertisers plan better.
  • Mentor at Scaler academy where mentored lots of Machine Learning enthusiasts on how to take their career forward and setup Machine Learning infrastructure, models and practices in their respective organizations.
  • Open Source contributor at numb3r33
  • Interested in competitive machine learning
    • Winner of Deep Learning Hackathon organized by HackerEarth (https://www.hackerearth.com/challenges/competitive/deep-learning-beginner-challenge/leaderboard/).
    • Ranked in Top-100 data scientists on Analytics Vidhya platform which has around 3500 competitors at the time of writing.
  • Interested in blogging about machine learning, deep learning and mathematics, you can read my posts here

This website is powered by fastpages 1.

  1. a blogging platform that natively supports Jupyter notebooks in addition to other formats. ↩