I am Tejas, a Senior Machine Learning Engineer at Swiggy, India’s largest and most valuable online food ordering and delivery platform. I am part of the data-science storefront, revenue-and-growth and trust-and-safety team solving problems such as homepage recommendations, ads personalization, fraud detection, intelligent discounting and smart notifications.
I have productionized large-scale machine learning models based on logistic regression, boosting, L2R, LSTM and BERT to accurately sprinkle personalization on top of tens of millions of daily visits in real-time. I also work closely with the platform team to build new features (recently: tensorflow serving) and involve in the design and solutioning of platform’s capabilities.
My current research interests are divided into two phases, first at the intersection of databases and ML (pre-modeling phase) to accelerate data exploration, data preprocessing and feature engineering and second, at the intersection of systems and ML (post-modeling phase) to build high-performance, scalable storage systems and model runtime that provide low latencies in high-throughput environment.
|Jun 14, 2020||Joining the location intelligence team to work on engineering and data problems in the geo-spatial domain|
|Apr 1, 2020||Promoted to Senior Machine Learning Engineer!|
|Mar 17, 2020||Our work on deploying deep learning models at scale at Swiggy using Tensorflow Serving is out of beta (primary contributor)|