AI copilot to slowdown your biological aging

Z fellowship for student entrepreneurs

Sigma Squared Society is a global community of the world’s most ambitious young founders selected for their impact-driven initiatives and their ability to transform broken industries.

Member of the inaugural On Deck health fellowship cohort

Building Brain Aging models via single-cell RNA Seq for human neurons

Working on developing Retrieval and Ranking models, ML model compression and distillation for Bing Search in Core Relevance team of AI&R

Building ML based strategies for high frequency trading in Indian equity and derivatives markets.

Working on fastening all-pair protein matching algorithms.

Extended the Karnik-Mendel type reduction algorithm to multidimensional space, so the type reduction could retain correlation among the features to an extent better than the traditional type reduction where each dimension is type reduced independently. I have proved theorems guaranteeing the optimality and convergence of the proposed algorithm. Likewise devised a multidimensional EIASC ( Enhanced KM algorithm). Worked on data preprocessing for improving the performance of neural nets by developing four methods of generation of Type-2 Fuzzy membership functions and proposed a new variant of KM algorithm for defuzzification and type reduction to use the resultant vectors as support vectors to transform the data into a new space.

Developed, implemented and analyzed methods for unsupervised classification of web-pages using deep neural networks. I had written the models using Microsoft's CNTK python.Worked on developing and implementing custom loss functions for deep neural networks for hierarchical classification problems

I have worked on devising and implementing a distributed hybrid bagging and boosted version of Gradient boosting machine for Decision trees training and achieved on par accuracy with baseline models with many factors decrease in computation time.