Mohak Patel

I work with Prof. Christian Franck to identify the mechanical force signatures of cancer cells. My research involves designing experiments and developing algorithms for 3D motion tracking and quantitative analysis of large image datasets to examine forces exerted by biological cells.

I am incredibly interested in AI and have built traditional machine learning and deep learning models for real-world applications in classification, computer vision, natural language processing, and recommendation. As a research project in computer science, I also developed a visualization for 3D stress tensor field in VR.

Outside of work, I travel frequently, enjoy football and click pictures (Milky Way).

I will defend my Ph.D. degree in early 2019. Currently, I am actively looking for full-time data scientist and/or machine learning engineer positions.

Research

Motion tracking algorithms

Rapid, Topology-based Particle Tracking for High-Resolution Measurements of Large Complex 3D Motion Fields

Mohak Patel, Susan E. Leggett, Alexander K. Landauer, Ian Y. Wong, and Christian Franck

Nature Scientific Reports 2018

A q-Factor-Based Digital Image Correlation Algorithm (qDIC) for Resolving Finite Deformations with Degenerate Speckle Patterns

Alexander K. Landauer, Mohak Patel, David. L. Henann, and Christian Franck

Experimental Mechanics 2018

Cell Mechanics

Mechanical Signatures of Metastatic Cancer Cells

Mohak Patel*, Susan E. Leggett*, Christian Franck, and Ian Y. Wong

Work in Progress

A cytoskeletal clutch mediates cellular force transmission in a soft, three-dimensional extracellular matrix

Leanna M. Owen, Arjun S. Adhikari, Mohak Patel, Peter Grimmer, Natascha Leijnse, Min Cheol Kim, Jacob Notbohm, Christian Franck, and Alexander R. Dunn

Molecular Biology of the Cell 2017

Paper

Scientific Visualization

Visualization of 3D Stress Tensor Fields using Superquadric Glyphs on Displacement Streamlines

Mohak Patel, and David H. Laidlaw

Manuscript Submitted

Projects

Data Science / Machine Learning

Best Vacation Planner

Web-scrapped, cleaned and analyzed the flight price, Airbnb accommodation, weather pattern and other data to plan the best and cheapest ski trip in Colorado. And predicted Paris Airbnb listing prices using tree-based boosting algorithms with a median absolute error of $20.12.

Project on Github

Microscope Image Cleaning

Developed a deep convolutional-autoencoder model in TensorFlow to remove noise and deconvolve microscope images. Generated synthetic images to train the model using theoretical optical model for a confocal microscope.

Project on Github

Tagging StackOverflow Questions

Developed a natural language processing model to predict tags for questions posted on StackOverflow. My best performing model was a deep convolutional multi-label sentence classification model implemented in TensorFlow.

Project on Github

Miscellaneous

3D Image Slice Viewer

3D image slice viewer in ipywidgets for Jupyter Notebooks, and Matlab.

Poisson-Disc-Sampling

Matlab script for N-dimensional Poisson-Disc Sampling

Project on Github