About

Swetha Revanur is a highly skilled machine learning professional with 3.77 years of relevant experience. She is currently working as a Data Products specialist at Brex, where she is leading a team to develop machine learning-enabled features for customers. Swetha has successfully shipped OCR-based bank statement processing, automated merchant categorization, and spend alerts, among other innovative products. Prior to joining Brex, Swetha worked as a Machine Learning Intern for Credit and Fraud at Branch International. There, she designed and built an LSTM-based fine-grained classification pipeline for SMS messages, achieving 99% accuracy for messages from previously seen senders and nearly 90% accuracy for messages from new senders. She also developed and integrated features to identify fraudulent borrowers and worked directly with global teams in Kenya, Nigeria, and Mexico. Swetha has also worked as a Machine Learning Intern at Amazon AI, where she designed and prototyped a new service for differentially private machine learning on federated datasets. This service had proven theoretical guarantees and achieved more than 20% higher accuracies in high-privacy regimes compared to state-of-the-art methods. Swetha met directly with customers and research teams across Amazon to identify requirements and was advised by Dr. Sudipta Sengupta. Swetha holds a Master of Science degree in Computer Science from Stanford University and a Bachelor of Science degree in Computer Science from the same institution. She was also the Valedictorian of Evergreen Valley High School. Swetha's technical skills include ML, AWS, Sagemaker, AI, Research Scientist, and Web. She is a highly motivated and results-driven individual with a passion for developing innovative machine learning solutions.