The mission of the Biohub at the Chan Zuckerberg Initiative is to empower scientists to cure disease.

In support of that mission, I collaborated with a microscope inventor to develop animations explaining Gradient Consensus, a novel algorithm for reconstructing true images from noisy scientific and medical data. For example, microscopes capture images that contain noise and blur, so Gradient Consensus reconstructs a sharper, more accurate image of the true object. Together, the animations and a written document I helped develop explained the algorithm's core concepts.

To ground the animations in real behavior, I ran the algorithm on synthetic test images in Python and captured how the reconstruction evolved at each iteration. Starting with a simulated cell image, I generated outputs showing how the algorithm progressively converges toward the underlying signal, distinguishes noise through gradient disagreement, and performs neighborhood-based updates when gradients agree. These computational results formed the foundation of the animations, ensuring they reflected the algorithm’s actual mechanics rather than illustrating the concept abstractly.