Swetha Subramanian     About     Research     Archive     CV     Portfolio

Swetha Subramanian

[Github] [Google Scholar] [Research Gate] [LinkedIn] [Bitbucket]



Hypothesis Testing, Machine Learning (Dimensionality Reduction, Random Forests, Regression, Classification, Neural Networks), Natural Language Processing (LDA, LSI, Cosine Similarity, Word2Vec), Statistical Analysis, Experimental Design, Neural Networks, Optimization, Signal and Image Processing, Control Theory, Numerical Simulations, Inverse Analysis, Computational Modeling, Kalman Filtering


Python (pandas, Matplotlib, numpy, NLTK, BeautifulSoup, PostgresSQL, Selenium, Gensim, scikit-learn), R, SQL, MATLAB, C, C++, Qt, Unix Shell Scripting (Bash), Git


  • PhD in Biomedical Engineering, University of Cincinnati, Ohio

  • Bachelor of Technology, Electrical Engineering, Jawaharlal Nehru Technological University, Hyderabad, India

Work Experience

Fellow, Insight Data Science, Palo Alto, CA

  • Developed a web application ConsciousChef (3-week project) that generates recipe recommendations to make the best use of a CSA box

  • Scraped recipes and produce shelf life information data from epicurious.com and other websites using selenium, urllib2, requests and BeautifulSoup which was then stored in postgresSQL

  • Developed an optimization algorithm to score recipes so that those featuring produce with shorter shelf life ranked higher

  • Deployed the web application using AWS, Flask, Bootstrap and jinja templating

Research Associate, University of Wisconsin, Madison, WI

  • Investigated the changes in the collagen structure of the cervix during pregnancy by quantifying changes in shear wave speeds using quantitative ultrasound techniques

  • Developed a GUI using Qt in C++ to perform a raster scan of cervix tissue in a 3D plane and obtain ultrasound data to generate high resolution 3D ultrasound images (< 60 micron) enabling us to understand morphological details of cervix tissue Git Repo

Research Assistant, University of Cincinnati, Cincinnati, OH

  • Investigated the utility of ultrasound echo decorrelation imaging method to predict thermal ablation during liver cancer therapy.

  • Probabilistically optimized computational finite element models of thermal ablation to generate accurate temperature profiles for comparison with echo decorrelation maps using a variation of unscented Kalman Filter. This resulted in temperature estimates within 5% experimental error

  • Estimated echo decorrelation thresholds for prediction of treatment boundaries and relevant temperature values with 80% specificity using ROC curves by comparing parametric echo decorrelation map with gross tissue histology and temperature maps

  • Published 5 papers in peer-reviewed journals

Awards and Honors

  • Biomedical Engineering Graduate Dissertation Award, Biomedical Engineering Department, university of Cincinnati

  • Best Student Paper Finalist, International Symposium of Therapeutic Ultrasound

  • University Research Council Summer Research Fellowship