Imperfect ImaGANation: Implications of GANs Exacerbating Biases on Facial Data Augmentation and Snapchat Selfie Lenses, arXiv 2020. This work highlights the negative impacts of mode collapse in Generative Adversarial Networks on social biases when trained on highly non-uniform real-world data. We train a DCGAN on images of engineering faculty in the United States and measure the differences in the original an generated distributions along race and gender. We also present a preliminary case study of impacts in a real-world application: we demonstrate that Snapchat's "female" selfie lens lightens skin tones in women of color.
#MeToo: Polarization and Discourse in the Digital Age, SocArXiv 2019. This work discusses how the introduction of the internet and social media inevitably impacts already-tense relations between characteristically different groups and illustrates how polarization impacts the digital vernacular. Curation causes insulation and polarization. This breeds inter-group rivalry, which in turn sparks outrage. I review literature on polarization and incivility in the internet and traditional media, and performs a case study on lay-people transmitting their views over Twitter in the #MeToo movement.
Graduate Research Assistant • August 2019 – Present Studying the impact of training set non-uniformity on mode collapse in GANs, and its implications for social media and web data.
Graduate Teaching Assistant • August 2019 – December 2019 TA for CSE 471: Introduction to Artificial Intelligence, emphasizing search, game trees, Markov Decision Processes, reinforcement learning, logic, neural networks, Bayesian inference, and classical planning.
Undergraduate Research Assistant • January 2018 – May 2019 Showcased inherent flaws in GAN-based data augmentation by training a model to generate faces of engineering researchers.
Undergraduate Teaching Assistant • August 2017 – May 2019 TA for CSE 310: Data Structures and Algorithms and CSE 340: Principles of Programming Languages.
Software Engineer Intern • May 2019 – August 2019 Provided advertisers optimal bid suggestions for cost-per-click ads by utilizing dynamic forecasted click-through rates customized to the campaign's objectives and targeting criteria.
Software Engineer Intern • May 2018 – August 2018 Implemented live preview in dynamic ads for businesses to see exactly what they are paying for and prevent duplication of effort and risk of inconsistency when advertisement designs change.
Student Partner • August 2016 – May 2019 • Contract Helped break the barrier of entry into tech by teaching workshops to engage students of all backgrounds and increase their exposure to up-and-coming fields, such as machine learning and the design and development of games, mobile applications, and websites.
Software Developer and Integrator Intern • August 2017 – May 2017 Built the Request Money feature for USAA members using Zelle, the revolutionary P2P payment service for instant transfers.
Subject Area Tutor • August 2016 – May 2019 Guided students on course concepts (focusing on foundational computer science and mathematics) in both individual and group settings and led exam review sessions.
Master of Science in Computer Science • 4.0 GPA • August 2019 – May 2020
Bachelor of Science in Computer Science • 4.0 GPA • August 2015 – May 2019 Outstanding Graduate for Computer Science, Spring 2019New American University Scholar: National Merit Finalist