Home/Blog/Alex Karpinsky
Picture of Alex Karpinsky

Alex Karpinsky

Alex Karpinsky (also known as homm on GitHub) specializes in high-performance image processing. Alex Karpinsky is best known for Pillow-SIMD, a SIMD-accelerated, drop-in replacement for Pillow that increases the speed of resampling and related operations.

Alex is a part of the Pillow core team and works at Uploadcare as an image processing engineer, where he authors technical deep-dives on topics such as fixing JPEG color banding by tuning quantisation tables and designing practical resize pipelines for modern CPUs. Beyond Pillow-SIMD, he develops libraries and benchmarks that bridge the gap between production requirements and low-level expertise, ranging from from HEIF/AVIF plugins and LUT tools for Pillow to ARM NEON and x86 AVX experiments.

Fun fact: Alex Karpinsky says he processed his first image in assembler at the age of 16.

Alex’s notable open-source work includes:

• Pillow-SIMD — a SIMD-optimized Pillow fork; drop-in compatible and measured 4-6× faster resampling vs. stock Pillow in official perf notes. Maintained on PyPI by homm.

• HEIF/AVIF plugin for Pillow — adds .heic/.avif read/write via pyheif/libheif.

• pillow-lut-tools — lookup-table loading/generation and creative filters for Pillow.

• pgauge — a CLI for monitoring CPU/GPU power usage on macOS using powermetrics.

• Pillow core — listed among release contributors and with ongoing participation in the ecosystem.

Expertise

  • Python; C and Assembly
  • SSE4, AVX2 (x86); SSE/AVX/AVX-512; ARM NEON
  • Pillow; Pillow-SIMD
  • JPEG; WebP; HEIF/AVIF
  • NumPy; OpenCV

Articles