Data Science, Machine learning, Computational Modeling

I like to solve problems through technology, engineering and data. I am a data scientist at Google and build data products to help Google Cloud Sales and Marketing teams bring revenue. Before starting at Google I worked for three years in medical device research and development at Boston Scientific. Before that, I got my Ph.D. in biomedical engineering where I focused on computational modeling and machine learning applied to biomechanical systems. Below are projects from my Ph.D involving computational modeling, multivariable regression and heuristic optimization applied to the human musculoskeletal system. I also present some side projects I did to learn and implement data science tools on fun datasets.

Ph.D. Projects

An overview of my Ph.D. research can be found in this slide deck on Slideshare.


Inference of analytical functions modeling finger movement using symbolic regression

I gathered experimental data consisting of finger movements and tendon excursions using a human cadaveric finger. Symbolic regression was implemented to infer the functions mapping joint angles to tendon excursions from this dataset. The form as well as the parameters of the mathematical function are simultaneously inferred from the experimental dataset.
Multivariable regression, machine learning.

Inference of anatomical finger models from sparse data

I implemented a stochastic hill climbing algorithm to optimize the structure and parameters of tendon networks. I used this to optimize the finger model to best match the experimental dataset that was collected and evaluated the accuracy of this model by predicting on a test data set. I used the Matlab Parallel Computing Toolbox and Distributed Computing Server to run the search on multiple processors.
Stochastic hill climbing, heuristic optimization

Data Science/Machine Learning Side-Projects

Predicting the rental count for a bike sharing company (Kaggle)

I performed exploratory data analysis using Python/Pandas/NumPy/matplotlib to understand which features were critical and gain more insight on the data. I then used Random Forest Regression in scikit-learn to predict the bike share count on the test data set.
Exploratory data analysis, Random forest regression

Predicting Survival on the Titanic (Kaggle)

This is a machine learning classification project based on a small dataset. I used this project to learn Python and the tools for data science (NumPy, scikit-learn, Pandas, matplotlib). I performed basic data cleaning and pre-processing. I used Random Forest classification to predict survival of passengers on the test set provided.

Contact information

The best way to reach me is through email.