GPU, CUDA, and PyTorch Performance Optimizations

Explore practical GPU, CUDA and PyTorch performance optimisation in an online AI Performance Engineering meetup on 21 September 2026 at 5:00 PM. This technical session focuses on improving the performance of AI workloads across the GPU software stack. Hosted by Chris F. for the AI Performance Engineering Meetup (London–Cambridge), it combines community updates with two talks dedicated to GPU, PyTorch and CUDA optimisation. ## What to expect The programme opens with introductions and meetup updates from Chris Fregly and Antje Barth. This is followed by two technical sessions, both listed under the theme “GPU, PyTorch, and CUDA Performance Optimizations.” Although the event page does not provide individual abstracts for these talks, the shared focus makes the meetup directly relevant to anyone interested in the systems and engineering work behind faster AI applications. The published agenda includes: - Talk #0: Introductions and Meetup Updates , Chris Fregly and Antje Barth - Talk #1: GPU, PyTorch, and CUDA Performance Optimizations - Talk #2: GPU, PyTorch, and CUDA Performance Optimizations The event takes place online using Zoom. It forms part of a recurring meetup series scheduled for the third Monday of each month through the end of 2026. ## Who it is for The technical subject matter is particularly suited to AI performance engineers, machine learning engineers, systems engineers, developers working with PyTorch and CUDA, and practitioners interested in GPU optimisation. It may also be useful for people exploring how underlying systems, algorithms and hardware affect the efficiency of AI workloads. No formal experience requirements are stated on the event page. ## Continue learning The organisers provide several supporting resources for attendees who want to explore the subject before or after the session:… Register for this event