Akash Haridas

Member of Technical Staff - Efficient Ai @AMD

Toronto, ON, CA

Systems Engineering- Maximizing distributed training throughput on large GPU clusters- Profiling and identifying bottlenecks in training workloads- Efficient kernels and custom fused operators- Supporting the training workloads of several large customers of our GPUsResearch on Efficient Architectures- Led the development of Nitro-T, an efficient image generation diffusion model trained from scratch in less than 24 hours- Core contributor to Nitro-1, a one-step distilled diffusion model for real-time image generation

Indian Institute of Technology, Madras

Dual Degree, Aerospace Engineering and Machine Learning

Paramekkavu Vidya Mandir, Thrissur

CBSE Class 12, Science

University of Toronto

Master's degree, Applied Computing

Warsaw University of Technology

Exchange Student