I consider myself a lifelong learner and believe in working hard. I graduated from UC San Diego with a Master's in Computer Science with a Machine Learning specialization in June 2018. I'm currently working at Twitter in the Timelines Quality team, building relevance and machine learning systems. Previously, I worked at Amazon with Amazon Expansions and Exports Tech team.
My interests are in the Applied ML, Recommender Systems, and NLP domains. I got my first exposure to Machine Learning during my internships at IIT Madras, India, developing scalable Bayesian models for the movie rating prediction tasks.
At UC San Diego, I worked on several research projects under the guidance of Prof. Julian McAuley. Recently, one of our research papers got accepted at ACL 2019. We proposed a neural network architecture to detect spoilers in review datasets and contributed an extensive book review dataset for spoiler detection. We published another work at RecSys 2018 last year. We proposed a new machine learning framework for predicting the fit of clothing products to recommend appropriate product sizes to customers. The proposed approach improves upon a method proposed by Amazon by up to 18%. In this work, we released two novel datasets, collected specifically for the task of clothing size recommendation.
Apart from this, I interned at Amazon in Summer 2017. My project was to support primary key constraints, batch inserts/updates, transactions, and compaction in Hive while ensuring read consistency. This project was particularly challenging as the Hive is not designed to handle transactional data and operations. I also worked with Arcesium, a recently formed subsidiary of D.E. Shaw & Co, as a Software Developer. I was responsible for revamping the infrastructure while ensuring scalability and robustness.
My industry experience helps me understand the intricacies involved in developing critical and time-intensive software. In contrast, my research experience is providing me an in-depth knowledge of the Machine Learning field. These learnings drive all of my work.