Siddhant Khadgi

Developed and implemented reinforcement learning models (e.g., multi-armed contextual bandits) to select optimal marketing email templates, using frameworks like Coach and Raylib. Deployed machine learning models on AWS SageMaker, leveraging SES for email distribution and creating webhooks for tracking interactions. Managed email templates and user data in MongoDB for efficient retrieval and personalized marketing. Implemented cross-domain user tracking via device fingerprinting, and created an OAuth 2.0 authentication server to process JWTs from device configuration and cookie data.

Worked in a startup which tries to digitize the retail market of India, by using computer vision and machine learning algorithms to predict what the customer is buying, so that the store can work autonomously. Used cameras with Raspberry Pi to collect real time data from retail stores. Used ARuCo markers to predict the location of a customer in the store. Used Microsoft Azure cloud services to collect video data and run computer vision algorithms.

Related Coursework: Database Management System and SQL, Data analysis and visualization, Machine Learning model training and inference.