This project focused on building lightweight, automation-adjacent tooling to measure real performance cost and validate rendering decisions. Rather than owning CI infrastructure, I created targeted systems that provided actionable metrics to reduce subjective judgment and inform engineering trade-offs.
Built Python-based performance test systems to capture GPU and runtime metrics
Profiled shaders, animations, and transitions directly on device
Supplied quantitative data to engineering and leadership to guide decisions
Identified performance regressions early and supported targeted optimizations
Integrated measurement into regular review and iteration workflows
Technical Artist
Python, PowerShell, GLSL (profiling context)
Instrumentation • metric capture • validation workflows • decision support
As visual fidelity increased, subjective evaluation became insufficient. This work introduced repeatable measurement to support informed decisions before and after public release.
