Ivan Pleshkov
Qdrant engineer. Led the implementation of TurboQuant vector quantization in Qdrant 1.18. Authored the companion engineering deep-dive blog post explaining every algorithmic extension: length renormalization, anisotropy compensation, P-Square calibration, support for L2 and unnormalized dot, and SIMD kernels.
Known For
- TurboQuant + RaBitQ hybrid implementation in Qdrant 1.18
- Anisotropy compensation via per-coordinate (shift, scale) calibration
- SIMD-optimized scoring kernels (AVX-VNNI, AVX-512, NEON)