Gilad Zamoscinski

Exploring strategies to improve performance in a diverse infrastructure, addressing issues with Skylake CPUs, AVX512 throttling, and TensorFlow updates, leading to notable enhancements in CPU frequency, load averages, and inference requests per core.

Learn how the Taboola team doubled the performance of their machine learning pipeline by optimizing the use of their GPU resources. Discover how they achieved increased throughput, reduced hardware costs, and improved recommendation engine AI.