Teaching Assistant

August 2020 - December 2020

CMSC 451 Design and Analysis of Algorithms
Conducted office hours and graded assignments and exams.
College Park, Maryland

Project Manager for Proofpoint, Inc. Clinic

August 2019 - May 2020

Malicious Message Clustering
Proofpoint is a security company based in Sunnyvale, California that provides software as a service and products for inbound email security, outbound data loss prevention, social media, mobile devices, digital risk, email encryption, electronic discovery, and email archiving. Our team focused on building an unsupervised retrieval framework using metadata associated with spamtrap messages to identify malicious emails of the same attack campaign.
Sunnyvale, California

Researcher in the Harvey Mudd Music Information Retrieval Lab

January 2019 - May 2019

Sheet Music Identification Using Measure-Based CNN Features
We presented a method for identifying a scanned page of sheet music by finding a match in a database of known musical scores. Our algorithm segments a page of sheet music into individual bars, extracts features from each bar using a pre-trained convolutional neural network with generalized- mean pooling, and then compares each bar feature vector in the query to the bar feature vectors in the database to determine the best match. To evaluate our method, we introduced a dataset containing 2,500 images of scanned piano music scraped from IMSLP.
Claremont, California

Machine Learning Intern at Unify ID

May 2018 - August 2018

Identification of Individuals from Smartphone Data
We designed and implemented an algorithm that identified individuals from smartphone accelerometer, magnetometer, and gyroscope data as a team of four. To do this we implemented functional PCA on the scipy spline class in python. We also implemented and fine tuned an algorithm to generate synthetic gyroscope data from accelerometer and magnetometer data.
San Francisco, California

Intern at Millenium Managment

May 2017 - August 2017

Stock Prediction Using Natural Language Processing on Earnings Calls
Designed and refined a NLP and Neural Net pipeline to scrape earnings calls and classify them as positive or negative for stock prediction over the period of a month.
New York City, New York