While completing a second MS in Neuroscience, I completed a research internship at CSHL to gain more hands-on research experience in an academic setting, which culminated in a publication in Nature Neuroscience: https://www.nature.com/articles/s41593-026-02262-8

• Created a 4D temporal-spatial recording system using multiple cameras, DeepLabCut, and Anipose to observe the dynamics of mouse-scale facial expressions.

• Built new recording hardware, conducted hands-on behavioral experiments with live mice, and organized team data collection activities

• Implemented software to automate analysis of experiments, enabling push-button results for other lab members

• Led ML anti-fraud program. Created research agenda, aligned plan with stakeholders

spanning multiple departments, and defined operational metrics to track success.

• Invented a novel metric to measure customer exposure to fraudulent profiles.

Created an intuitive visualization by implementing a “Nowcast” model to extrapolate

present-day values from lagging indicators, enabling company-wide visibility.

• Created an extensible “Early Warning System” to monitor and send alarm messages

concerning model drift, data quality, and other operational issues.

The Fashion Innovations organization invents new ways for Amazon to sell clothing.

• Launched two new customer experiences: the Made For You product line (Amazon’s entry into the online made-to-measure space) and Virtual Try-On (a 3D visualization of how the garment would fit on a customer’s body).

• Created the Body Estimation System, which predicts clothing sizing using computer vision and machine learning.

• Led team operational-readiness activities, including load testing, alarming, and bug triage. Coordinated efforts of 4 other engineers, resulting in a 100x increase to maximum concurrent traffic in the three months prior to launch. Surpassed availability goal and had zero operational issues during launch window.

• Introduced Jupyter notebooks as a lightweight alternative for developing internal tools. Enabled 7 engineers across 4 teams to collaborate on cross-functional projects.

• Launched Athena Web project, which enabled customers to access pre-existing

desktop applications in a web browser.

• Built a high-performance javascript implementation of a grid widget, which is extremely configurable and capable of displaying billions of rows of tabular data

• Integrated web controls to the charting API, allowing customers to interact with server-rendered charts. I implemented the Python Tornado back-end service and used websockets for client-service communication.