Sep 2022 – Jul 2024 · 1 yr 10 mos · Singapore

Jun 2019 – Jun 2022 · 3 yrs · Singapore

May 2017 – May 2019 · 2 yrs · Bengaluru Area, India

May 2016 – Jul 2016 · 2 mos · Bangalore

• We explored new ways to improve YouTube user profiles and video/channel recommendation by modelling temporal variation of user interests. We learnt sequence relations among YouTube channels/videos based on their consumption by a single user. We used Bayesian method and neural networks(LSTMs and artificial neural networks using Tensorflow) to learn the video/channel transitions. We used distributive frameworks like MapReduce to deal with large amount of data.

May 2015 – Jul 2015 · 2 mos · Bangalore

• We created first-of-its kind tool to summarize a multimedia article that captures information from text and image segments. We used techniques like Deep Learning based language processing and computer vision to create the tool. We also conceptualized a novel metric evaluating article summary taking in account cohesion between image and text.
• We published paper "Summarizing Multimedia Content" at WISE, 2016.
• We also filed two patents “Multimedia Document Summarization” and “Evaluation of Multimedia Summarization” to the US Patent Attorney, after internal acceptance at Adobe San Jose

Jan 2015 – Apr 2015 · 3 mos · IIT Kharagpur

• I worked for Flipkart as an extern on the project "information cascade in the e-commerce domain". We analyzed market cascades on Yelp for products based on Yelp for products based on the propagation of information in a geographical region. We showed that "outsiders"(non-local consumers) to business units are responsible in bringing consensus of choice as compared to "insiders"(local consumers) who are considered as experts of the respective local business units.
• We published a poster “The role of outsiders in consensus formation: A case study of Yelp” in the prestigious ACM CSCW, 2016.