New Brunswick, New Jersey, United States
Currently, I am a software engineer at Bloomberg LP, on a team working on Bloomberg's tools for work-from-home. Previously, I was a master's student in computer science at Princeton, where I did research on machine learning. I was involved with a variety of groups on-campus, including the New Jersey Student Climate Advocates, Princeton Effective Altruism, and Princeton Robotics. I'm always down to chat about machine learning, infosec, climate solutions, or effective altruism.
Developed and supported Bloomberg’s remote access tools. The main products I support are Bloomberg Anywhere, our company's flagship remote option for accessing the Bloomberg Terminal, and the Bloomberg employee remote access portal, which facilitates all employee remote access to Bloomberg corporate systems.
Conducted research on performance optimizations for deep learning Topics of research: reducing memory usage of gradient descent by compressing momentum, and reducing data I/O costs with data echoing Designed and executed experiments in GCP Compute Engine and AWS EC2 Analyzed algorithm performance on training logistic regression on RCV-1, ResNet on MNIST, GLoVE on the 1 Billion Words Dataset, and similar
Migrated Information Security course infrastructure, a mini network of three machines for students to hack, from DigitalOcean to AWS CDK Tools used: AWS CDK, Route 53, EC2, S3, Let’s Encrypt Supervised graders, created rubrics, taught precepts, and held office hours Commended by name in 31% of Spring 2020 Intro to ML course reviews
• Designed and built a user interface to monitor usage of AWS VM’s, using Sketch, React, and PostgreSQL • Collaborated with a remote team to migrate a data pipeline segment from a SQL script to Java
In the Cohen Lab, I contributed to PsyNeuLink, an open-source Python toolkit which simplifies computational neuroscience modeling • Designed and implemented feature to automatically adjust inputs to single-layer recurrent networks in PsyNeuLink • Collaborated to build, document, & maintain PsyNeuLink interfaces to Emergent, PyTorch models
• Designed and built a full-stack application in Scala, C#, and React, used by company tech support staff to query for changes to customer settings database • Gathered and applied user feedback to improve on application after it was shipped to production • Analyzed performance, optimized to improve performance on large data • Documented, refactored, and explicated code design to facilitate code maintenance after departure