Jordan Graves

A review paper cataloging various deep learning techniques used in antibody engineering. I wrote this paper with help from colleagues while serving as the head machine learning expert at Macromoltek in 2020.

AstroNinja was a personal project of mine which included a back-end powered by Python and the Google App Engine and a front-end built using C# and Unity3D. My favorite features are the smooth blended animations using the Mecanim system and the audio-driven animations using the real-time FFTs of audio loops. Since the audio was such an important part of the gameplay experience, I decided to compose the music and create all of the audio effects from a few instrument libraries myself. The result is a game which intimately blends visuals, audio, and gameplay.

ReadZinc was an avant-garde idea which offered a new way to consume text-based media. At the heart of the project was a method to procedurally generate art and text-geometry using passage context and grammar elements. The front-end application was built using one of my favorite platforms, Unity3D. The application was very graphic intensive. Since the application was designed for mobile devices, we worked at the limit of what the hardware would allow.

A client needed an Android application to be ported to iOS in a short time frame. The application was very UI intensive and had several features including user-user and user-group messaging, location tracking, location triggers, and several others features. Although originally designed for Android, the UI and application, in general, functioned well and looked great!

I worked on a team of no more than five people to build a mobile application to support a startup as fast as possible. I implemented both the iOS client as well as all of the server-side code. Our application supported multiple user types and allowed for scheduling deliveries and pickups as well as routing information to couriers.

Simulations of soft tissues rely on constitutive models whose form must be generated by in-vitro experimentation. For membranes and thin specimens, planar biaxial testing systems have been used, but have always been limited in their ability to fully prescribe in-plane F2D. We developed a novel planar biaxial testing device that was capable of full control of the in-plane deformation gradient tensor F2D and of testing specimens as small as ~4 mm x 4 mm. A sliding motion controller implementation drove 12 independent actuators arms and allowed us to enforce any arbitrary F2D with a high degree of accuracy.

I worked at multiple levels on this project: writing analysis software to assess device performance, determine control parameters and analyze final device results. I even designed some of the hardware components in SolidWorks. The analysis software was written in MATLAB and Mathematica and the UI and control software was written in LabVIEW.