Hi there, I am a computer graphics / video games software engineer at Deviation Games. Dabbler in rendering, ML, game design, and occasionally homebrewing table top stories βοΈ.
Previously, I worked on neural rendering research at Google Stadia R&D and also AI performance for Red Dead Redepmtion 2 at Rockstar Games.
I recently resleased the SyntheticFur dataset in collaboration with Google Brain and Stadia.
You can browse the sidebar for some of the things I that have worked on.
Designed and implemented real-time rendering features in Unreal for an unannounced AAA title.
Advocated for and applied neural rendering research toward solving real-time graphics problems for video games. My focuses were AI art generation and high fidelity fur rendering. Worked on generating custom 3D datasets using Houdini and Blender, training generative machine learning models with Google infrastructure, parameter tuning, and interfacing novel workflows with artists.
Analyzed and optimized performance for AI systems in Red Dead Redemption II, Rockstarβs premier narrative-driven, open world video game title with deep AI interactions. Specifically worked to improve the C++ game code for LOD management of AI physics, weapon and combat systems, transport systems, and pathfinding systems for PS4 and XBox. Worked on general AI and animation optimization for several specific scripted in-game missions as well as miscellaneous performance bugs in the final weeks before release.
Implemented VR editing tools for Unreal Engine 4. The tools include terrain editing, virtual keyboards, and foliage painting for level editing. These tools were introduced in the Unreal 4.13 and 4.14 updates.
Developed infrastructure and applications for the UP3 fitness wristband on ARM Cortex and iOS platforms. This included the BTLE protocol, authentication and encryption between devices and the mobile app, activity classification collection tools, peripheral drivers, USB interface, and UX.
Designed a GUI with Python QT for the systems used in Wireless Resonant Energy Link (WREL) research. The software supports data collection, data visualization, wireless control, and power diagnostics. Over the years, this software has been extended for use at the startup company Wibotic and in other research ventures.