Fast & Robust Divide-and-Conquer 2D Delaunay Triangulation in C99
How to construct $O(n \log n)$ 2D Delaunay triangulations, guarantee numerical robustness with adaptive predicates, and optimize quad-edge topologies for CPU cache lines.
System Architect, High-Performance Computing specialist, and low-level software engineer. Pushing the limits across AI, Distributed Systems, Computer Graphics, and Real-Time Computing.
Selected libraries and systems spanning computational geometry, GUI frameworks, parsers, and low-level performance utilities.
Contributions to computational geometry, $k$-set triangulations, fingerprint ridge frequency estimation, and real-time graphics.
Wael El Oraiby, Dominique Schmitt, Jean-Claude Spehner • Applied Delaunay triangulation structures to enhance low-quality fingerprint image ridge frequency analysis.
Wael El Oraiby, Dominique Schmitt, Jean-Claude Spehner • Investigating higher-order centroid triangulations derived from planar point sets and dual arrangements.
Wael El Oraiby • Doctoral thesis detailing combinatorial geometric properties, divide and conquer algorithms, and dual structures of order-$k$ inclusion chains.
Wael El Oraiby, Dominique Schmitt, Jean-Claude Spehner • Analysis of topological inclusion chains and order-$k$ Delaunay triangulations in $\mathbb{R}^2$.
Wael El Oraiby, Dominique Schmitt, Jean-Claude Spehner • Efficient geometric divide and conquer algorithms for polygon boundaries of $k$-sets.
Dave Astle et al. • Author/Contributor covering advanced OpenGL rendering techniques, shader architecture, and real-time gaming algorithms.
Dated technical deep-dives into algorithms, Rust/C internals, and system design.
How to construct $O(n \log n)$ 2D Delaunay triangulations, guarantee numerical robustness with adaptive predicates, and optimize quad-edge topologies for CPU cache lines.
How we ported rxi's microui to Rust in microui-rs. Tackling borrow checker constraints without resorting to raw pointer soup, and architecting command buffers with zero heap allocations during the render loop.
A systems war story: Scaling an online advertising pacer to 5 million bid requests per second with sub-2ms latency by combining F#, CoreCLR, libuv in C11, and the Actor model.
Senior technologist blending mathematical rigor with high-performance low-level systems execution.
With extensive experience across tier-1 technology leaders (NVIDIA, Microsoft, Samsung) and leading high-throughput systems at NeoCogi, my primary work revolves around:
If you have questions, suggestions or your company has an interesting and challenging project, don't hesitate to contact me on: