Arpitha Srinivas is an AI Systems Performance Engineer at Supermicro, specializing in optimizing large-scale AI models such as Llama 4, Llama 3.1, Llama 2-70B, BERT, Stable Diffusion, ResNet, and 3D-Unet across NVIDIA (B300, B200, H200, GH200, H100), AMD (MI350, MI325X, MI300X), and Intel (Xeon) platforms. With a master’s degree in AI Software Engineering from San José State University, she brings deep expertise in TensorRT, vLLM, CUDA, ROCm, and MLPerf benchmarking. She has successfully submitted multiple rounds of MLCommons MLPerf Inference and Training benchmarks. Her research paper, "Cyber-Security Dashboard: An Extensible Intrusion Detection System for Distributed Control Systems," was accepted at the IEEE SVCC 2025 Conference. Arpitha is a passionate advocate for women in technology and has contributed to several industry conferences. She also reviews for O’Reilly Media books. When not tuning AI systems and models she enjoys hiking, music, and painting.
