I studied computer science and cognitive science at Rutgers university, at which I was a presidential scholar (top 2% of class). After interning at Amazon, I developed an interest in NLP and learned how AI can be used to solve problems in healthcare. I currently work on the Amazon Translate team where I work with research scientists to develop and host scalable neural machine translation (NMT) systems for 70+ languages across 16 regions worldwide. .
During my free time, I enjoy volunteering in settings where I can apply my technical knowledge to help others. Over the past few years, I have learned about various industries by volunteering at research labs, political campaigns and data-centric nonprofits, such as DataKind. I am passionate about using machine learning and data science to solve social problems. In particular, I hope to address issue related to the lack of interpretability in machine learning models and how this leads to the propagation of unseen biases.
Nov 2022 - Present I'm excited to announce that I have joined NextPatient! NextPatient is a patient scheduling and engagement service. I worked directly with CEO to own the company's engineering and product development. Over the past year I have worked on revitalizeing the service's frontend (HTML/CSS), upgraded backend services (Python/Flask), and updating the data model (SQL/Postgres). During such time, topline revenue has increased by 33%.
Aug 2020 - Nov 2022: For two years I worked as a part of the Amazon Translate team. Amazon Translate is a neural machine translation service that processes large amounts of multilingual data and allows you to use real-time and asynchronous translation via a simple API call.
My contributions to the team include:
- Spearheading re-architecture of Amazon Translate’s ML model hosting to reduce costs by 25%. Specifically, I devised new routing logic for single model to multi-model per-host migration.
- Redesigning internal translation website using Node.js, Material UI design standards, AWS CDK, and AWS Cloudfront
- Engaging in various ML Ops improvements by improving the interpretability of ML models via reporting and accelerating model release pipeline
Summer of 2019: I returned to Amazon but I will be joining the Comprehend Medical team. My new team works on using NLP to parse patient medical records and drastically expedite medical research. In doing so, we automate the extraction of information and will empower researchers to collect and analyze data at scale, and mitigate the computational bottleneck of research.
Summer of 2018: I worked in Seattle as a software engineering intern within Amazon's Aurora Database team. As a member of one of the largest cloud database business, I learned how systems are built to manage a huge numbers of databases and vast amounts of storage across multiple datacenters worldwide. To achieve this level of reliability, my team builds control and monitoring systems that automatically detect and handle many types of failures within seconds, and data replication options that accommodate various geographical distribution and disaster recovery objectives. I built a tool using AWS ElasticSearch to parse unstructured log data and allow developers to search for errors using ad hoc queries.
DataKind: I volunteered as a Data Expert to help generate insights regarding housing insecurity in Central Florida. My teams goal was to generate insights as to the equitable distribution of emergency rental assistance program funds. We partnered with the Housd Eviction & Foreclosure Group to help more residents of Central Florida to stay in their homes.
Shor's Quantum Integer Factorization Algorithm: Many modern encryption algorithms (including RSA) rely on the assumption that factoring large integers is computationally intractable. This is true to classical computers, but the Shor's algorithm shows that factoring integers is efficient on an ideal quantum computer, so it may be feasible to defeat RSA by constructing a large quantum computer. I have implemented a Shor's algorithm using IBM's Qiskit and have ran experiments on small quantum computers.
Shared Autonomous Vehicles Simulation: a simulation of Shared Autonomous Vehicle (SAV) Deployment. I analyzed the efficacy of using electric vehicles in NY as well as locating the optimal charging stations based on historical taxi data. I mainly worked on creating the simulation as well as the data analysis for this project.
#Farewell Obama: During Obama's farewell address, I scraped tweets from Twitter and conducted sentiment analysis (using IBM Alchemy) and network analysis of all tweets with the keyword #Obama
Find my thoughts on Twitter, my projects on my Devpost, and my code on GitHub.
Feel free to reach out by sending me an Email or message me on LinkedIn.