Developing a platform-agnostic prefix-sum algorithm based on Nvidia research. Employing novel kernel optimizations, such as parallelized lookback and threadgroup data bit-packing. I demonstrated that the latter optimization increases GPU throughput by an average of 20 GB/s by batching Vulkan atomic API calls. This work will be featured in an upcoming article.Conducted revisions to our open-source Vulkan simplification layer (Easyvk), refactoring our buffer allocation process, implementing buffer staging, in order to add an option for device-local buffers.

Not So Slow Slugs: High Performance Computing Team @ UC Santa Cruz.Competing in the IndySCC 2024.

Satellite design and fabrication clubhttps://slugsat.soe.ucsc.edu/Contributing to the Pleiades Project as one SlugSat's Club Leads, after our joint proposal was accepted by NASA’s CSLI program. Set to develop a C++ wrapper using NASA's F´ framework for CubeSat flight software as part of the PROVES flight framework.

Accomplishments:- Fastest 100m in program history. Achieved at 2024 Hornet Invite (10.83).- Furthest Long Jump in program history. Achieved at 2024 Chico Invitational (22’4).