Shishir Sanjeev

Developed a model for forecasting financial KPIs in the hospitality industry, using location, ratings, geospatial variables, and hotel classifications (3/4/5-star) to account for market dynamics. I cleaned raw data, used krigging, created a hierarchical machine learning pipeline for time series forecasting to achieve 80-85% accuracy with real-world data.

In my currently ongoing master’s thesis, I combine techniques from stochastic processes, topological data analysis,]convolutional neural networks, and global optimization techniques for the purposes of classification of protein structures. This has significant applications for drug discovery and design.