OpenCell is a collaborative project led by the Leonetti group at the Biohub to map the architecture of the human cell by first building proteome-scale libraries of tagged human cell lines, then using these libraries to systematically measure protein interactions and localization patterns. I was the sole developer of opencell.czbiohub.org, a full-stack web application to manage and visualize the large multimodal datasets generated by this effort. I was also a co-first author on the paper we published in 2022 describing this project.

This was a project at the Biohub to automate the imaging of 96-well plates by combining microscope-control logic with real-time predictions from a pre-trained machine learning model to determine optimal imaging positions.

This was a side project to build intuition for how classification models work. It trains models directly in the browser using plain JavaScript implementations of common classification algorithms (including SVMs and neural networks), and renders a dynamic real-time visualization of how model predictions fluctuate during training.

This was a side project to interactively visualize a collection of century-long daily temperature records from stations across the US. It was primarily an exercise in using d3.js to build bespoke visualization dashboards.