Kaggle Bronze Medal - Handwritten Bengali Character Classification
Created computer vision models using PyTorch to classify components of 200k+ images of Bengali characters. Received a Bronze medal for placing in the Top 10% of competitors.
Hi, my name is Ash.
I am a Data Scientist at Retrace Labs, a startup in the Bay Area. At Retrace I am responsible for developing and deploying machine learning models using Python to automate the dental claim adjudication process. Previously I was a Senior Data Analyst at Canadian Tire. I have a Master's degree in Data Science from the University of San Francisco and a Bachelor's Degree in Industrial Engineering from the University of Toronto. I have over 5+ years work experience using Python (pandas, numpy, sklearn, PyTorch, Flask), SQL, Docker, AWS and other tools to solve data science problems.
I'm a big fan of anything to do with basketball, horror movies, rap, indie, and pop music, and travelling.
My Work Experience
| Role | Company | Location | Dates |
|---|---|---|---|
| Data Scientist | Retrace Labs | San Francisco, CA | Oct 2019 - Present |
| Senior Data Analyst | Canadian Tire | Toronto, Canada | Sep 2015 - May 2019 |
| Business Operations Analyst, Intern | IBM | Toronto, Canada | May 2013 - Sep 2014 |
My Skill Set
I have extensive experience using Python for machine learning and deep learning projects. I've worked on all parts of the data science workflow, including requirements gathering, data acquisition and cleaning, modelling, deploying, and communicating results to stakeholders.
Projects
Outbreak Project (Private Repo) - Interviewed by YCombinator
Created models to predict flu outbreaks using social media data. Achieved 91% R squared by creating features from Instagram hashtag and image data, as well as from historical flu trend data. Interviewed by the non-profit arm of YCombinator.
Indigo Music Studio
Built on top of Google Magenta, we developed a web app that allows users to harness the power of machine learning models to create their own music. Users can interpolate between 2 tracks or generate accompanying instrumental tracks.
Seattle Public Library System Book Check Out Predictions
Beginning with over 35GB of data, we predicted which books will be checked out over the next month using PySpark, Amazon EMR, and H2O.
Is Pitchfork the "Voice of the People"?
Evaluated the relationship between the commercial success of an album and the score it receives from Pitchfork by creating visualizations in ggplot and Bokeh.