Rachit Garg

Rachit Garg is a deep learning architect at NVIDIA, where he focuses on advancing the performance and efficiency of large-scale AI training. In recent years, his work has centered on improving the training of state-of-the-art foundational models and MLPerf benchmarks through innovations spanning algorithms, software, and hardware. His work now extends from optimizing production AI workloads to influencing future NVIDIA platforms.
Avatar photo

Posts by Rachit Garg

Agentic AI / Generative AI

Boosting MoE Training Throughput with Advanced Fusion Kernels

Mixture-of-experts (MoE) models have quickly become a foundational component of modern, large-scale AI systems. They are widely adopted because they enable... 9 MIN READ