"In God we trust; all others bring data."
-- W. Edwards Deming

About Me

I am a full stack data scientist with over 8 years of experience solving business problems using machine learning and experimentation. I graduated from IIT Kharagpur in 2014 and worked as a Data Scientist / Assistant Manager at Loylty Rewardz for 5 years. Here, I deployed machine learning algorithms to help our banking clients understand the behavior of their card holding customers and implemented strategies to increase customer stickiness.

I recently completed my Master's degree in Data Science from the University of San Francisco with a GPA of 3.97 (top 5% of my class). I am currently working at Shipt as a Senior Data Scientst where I am optimizing gig-worker pay by building fraud detection framework.

Scroll down to check out some of my interesting projects.

Machine Learning - Real World Issues

Predicting Library Checkouts

A classification model to predict book checkouts next month.

(PySpark, AWS EMR, H2O)

Energy Consumption Prediction

A regression model to predict energy usage in metered buildings.
(Scikit-learn Pipelines, Randomized Hyperparameter Search)

Vehicular Traffic Vs Economy

An analysis of trends in traffic volume and economic activity.

(Python, ggplot, xlrd)

Cool Applications / Publications

Math2Code.com

Free application that converts pics of math equations into Python code.

SlateQ

Google's Reinforcement Learning algorithm for recommender systems.

Python Implementation - ML Algorithms

Random Forests

A recursive approach to building random forests from scratch.

K-means Clustering

K-means clustering with kmeans++ centroid initialization algorithm.

Feature Importance

An exploration of common feature importance and selection techniques.

Decision Trees

A recursive approach to building binary decision trees.

Regression

Linear, Ridge and Logistic Regression using AdaGrad optimization algorithm.

Naive Bayes Classifier

A multinomial classifier to predict sentiment of movie reviews.

Hobbies

Check out my killer serve.

Stay Connected