2026 withby Gabriel
Worked with Christopher's Kitchen, a 501c3 non-profit, to automate their donation value tracking process. Built the process through a phone barcode scanner, a UPC database, Google Sheets, and Google Apps Scripts. Definitely my most visibly impactful project to date, we were able to free volunteer hours by reducing the time spent tallying donations by over 75%.
Developed the Live Sequence Protocol (LSP), a network protocol combining ideas from TCP and UDP to support a client-server communication model. The network protocol included techniques like the sliding window algorithm, acknowledgements, heartbeats, and checksums. Using this, I then built a load-balancer to distribute 'bitcoin mining' load across multiple of these custom servers.
Contributed to Mermaid.js, an open-source text-based diagramming library used for technical documentation. Made PR's for: fixes to Gantt diagram duration parsing, support for accented characters in chart labels, and configuration bugs in gitGraph that prevented dynamic updates.
Super interesting project taking a deep dive on the Raft distributed consensus algorithms. Using the paper by Diego Ongaro and John Ousterhout as a guide, I built a version from scratch in Go supporting leader elections, log replication, and the long list of edge cases associated with maintainence across network partitions and node failures.
My own little nook of the internet. This was a super fun way to get more versed in web dev and exercise my creative muscles. I developed it using Next.js and a lot of inspiration from many other beautiful websites. After spending many hours staring at pixels until my eyes were bleeding, I have a newfound respect for designers.
Implemented a neural network in Python to perform handwritten character recognition on an OCR dataset. The project involved deriving and coding forward propagation and backpropagation by hand, implementing sigmoid and softmax layers, and training the model using stochastic gradient descent.
