Apr 2021 – Oct 2023 · 2 yrs 6 mos

• Led ML initiatives across storefront including the food page, menu, payments, and checkout. Our team built and maintained high impact, high scale Machine Learning models for restaurant listing, dish recommendations and other high visibility storefront properties.

Apr 2019 – Mar 2021 · 1 yr 11 mos

• Worked on personalization of the menu page (item ranking) and optimisation of payment routing models.

Apr 2017 – Jan 2019 · 1 yr 9 mos

• Led the Data Science team for Customer Experience. Solved problems in the areas of
• escalation prediction, returns fraud, customer churn, and others. My role required me to wear
• multiple hats - building models, translating business problems into data science solutions, guiding the team with approach, interacting with business and product stakeholders.

Jan 2016 – Mar 2017 · 1 yr 2 mos

• Built both predictive and deterministic models for Customer Segmentation and Customer Churn Prediction respectively. These models power Customer Experience tools that are used by CC agents to offer differentiated services to customers.

Sep 2014 – Dec 2015 · 1 yr 3 mos

• Built short term intent prediction models. To be specific, given a user's browsing data the models predict if s/he is showing intent to purchase and what style(s) will be of interest to him or her. These models powered several use-cases like personalised recommendation emails, personalised notifications and personalised coupons.

Sep 2012 – Aug 2014 · 1 yr 11 mos · Bangalore, India

• As a part of the Smarter Education group, I worked on an umbrella initiative called Personalized Education Through Analytics on Learning Systems (PETALS). I led the effort of building predictive models to detect students who are at risk of performing poorly and are not college-ready. This detection helps teachers to intervene and correct the student’s learning path. The system is currently successfully deployed in one of the largest public school counties in the US. My work at IBM Research contributed to one journal paper, two conference papers and three patents (two filed, one published).

Jun 2011 – Sep 2012 · 1 yr 3 mos

• Wrote the backend for a system called 'Autopoint'. Autopoint is capable of automatically generating powerpoint slides from a corpus. There were several sub-problems in this project - my prominent contribution being writing the algorithms for multi-document summarisation and sentence shortening (creating bullets from sentences). This work contributed to filing of a patent and publishing of a tech report.

Jun 2010 – Aug 2010 · 2 mos

• Worked on implementation of Video Summarisation algorithm that given a video, it's text transcript and expected duration of video, summarises the original video keeping the gist intact. The approach was to summarise the text transcript of the video and then chop the original video. Abrupt visual changes in the summarised video were avoided using heuristics on difference in video frames. This work contributed to publishing of a tech report.

May 2010 – May 2010 · 0 mo · Pune

• Short internship to study different data mining algorithms and their parallel versions. Used Cilk++ to implement parallel version of A-priori algorithm.