I design controlled experimental frameworks to isolate CPU/GPU bottlenecks and protect deterministic frame budgets in real-time systems.
Real-time performance is not optimized by guesswork.
It is engineered through:
- Deterministic execution models
- Controlled stress harnesses
- Reproducible profiling experiments
- Frame-time variance isolation
- Systems-level bottleneck diagnosis
Every project below reflects that philosophy.
My current work concentrates on:
- XR frame stability under stereo rendering constraints
- Overdraw amplification in tile-based GPU architectures
- Deterministic frame-budget control under dynamic scene complexity
- CPU/GPU bottleneck isolation in real-time pipelines
- Telemetry-driven runtime health monitoring
The objective is not higher peak FPS, but predictable frame delivery under stress.
Deterministic XR rendering stress framework for isolating GPU bottlenecks, overdraw amplification, and frame-time instability under controlled load.
๐ https://github.com/JamesDeRaja/XRPerformanceLab
Empirical performance analysis of Unity UI soft masking under layered transparency and controlled overdraw scenarios.
๐ https://github.com/JamesDeRaja/SoftMaskPro-Performance-Study
Performance engineering is not about raising FPS.
It is about:
- Frame budgets defined in milliseconds
- Variance isolation
- Non-determinism elimination
- Telemetry-first system design
Architecture > Optimizations
Measurement > Assumptions
Determinism > Peak numbers
Full case studies and extended technical breakdowns:
๐ https://james.alphaden.club
LinkedIn: https://www.linkedin.com/in/james-de-raja/

