I specialize in high-performance computer vision and systems programming, working primarily with Rust and Python (PyTorch). My focus is on removing bottlenecks, optimizing memory layouts, and making complex algorithms viable for resource-constrained environments.

Bloomsday is a very high-performance, cache-line blocked Bloom filter designed to be the fastest option for modern CPUs.

• Unmatched Speed: Benchmarks show ~2.3x faster lookups than other optimized crates like sbbf-rs by leveraging compiler-friendly loops for auto-vectorization (SIMD).
• 100% Safe & Lightweight: Achieves top-tier throughput with zero unsafe blocks and zero heavy dependencies.

As a maintainer of the Kornia ecosystem, I focus on massive speedups and memory reductions.

• Memory-Efficient AprilTags: Reduced core data structure memory from >6GB to ~450KB (14,000x reduction), enabling edge deployment.
• High-Speed Retrieval (kornia-bow): Authored a pure-Rust Bag of Words crate with 11x faster lookups and 5x lower memory usage vs. reference implementations.
• PyTorch Optimizations: Achieved 50x–300x speedups in core operations (Otsu thresholding, geometric estimation, 3D masking) via vectorization and analytical solvers.

I'm always looking for interesting real-world optimization problems. If you have a bottleneck you're trying to crush, feel free to reach out and we can tackle it together.

Languages
Frameworks & Libraries
Databases & Cloud
Tools