We’re excited to announce that PyTorch will have a strong presence at NVIDIA GTC 2026, from March 16-19, 2026 in San Jose! Whether you’re a seasoned PyTorch developer or just getting started, we invite you to join us for demos, talks, hands-on labs, and opportunities to connect with PyTorch core maintainers and community experts.
Visit the Meta Booth — Booth #338
Meta will be presenting a booth with to PyTorch. Stop by Booth #338 to meet PyTorch experts from Meta and the PyTorch Foundation. We’ll be showcasing:
- Helion Demo: Get hands-on experience with Helion, a new PyTorch-native kernel authoring framework. Learn how to write custom kernels and autotune them for optimal performance on your target NVIDIA GPU.
- ExecuTorch: See NVIDIA’s Parakeet speech-to-text model deployed using ExecuTorch for high-performance, Python-free, on-device inference, demonstrating an end-to-end edge pipeline from audio capture to transcription.
Featured Talk: From Kernels to Clusters — How PyTorch Powers High-Performance AI
On Monday, March 16, from 4:40 to 4:55 PM PT, join Alban Desmaison (PyTorch core maintainer at Meta) for “From Kernels to Clusters: How PyTorch Powers High-Performance AI” at GTC 2026.
In this featured GTC session, Alban Desmaison will walk through:
- A broad overview of the PyTorch framework and its evolving ecosystem
- How Meta and the broader community leverage PyTorch for high-performance and distributed AI workloads
- Key performance wins and updates powering state-of-the-art research and products
- What’s next for PyTorch: insights into the technical roadmap and upcoming features
Add this talk to your NVIDIA GTC agenda to gain a firsthand perspective from the PyTorch core team!
PyTorch Sessions & Talks
Monday, March 16
| Time (PT) | Session | Speakers |
|---|---|---|
| 3:30 PM | Open source RL and Agents, highlighting TorchForge and OpenEnv (CoreWeave Booth) | Hamid Shojanazeri (Meta)
Aaron Batilo (CoreWeave) |
| 4:40 PM | From Kernels to Clusters: How PyTorch Powers High-Performance AI | Alban Desmaison (Meta) |
Tuesday, March 17
| Time (PT) | Session | Speakers |
|---|---|---|
| 1:00 PM – 1:50 PM | Turbocharge Your LLM Inference: Expert Strategies for Lightning-Fast, Scalable AI Deployment | NVIDIA Experts |
Wednesday, March 18
| Time (PT) | Session | Speakers |
|---|---|---|
| 3:00 PM – 4:45 PM | Optimize PyTorch Models for High-Performance Inference With Nsight Deep Learning Designer (Hands-on Lab) | Manoj Kumar Yennapureddy (NVIDIA) |
| 4:00 PM – 4:40 PM | Build Fault-Tolerant Distributed AI Training at Scale | Shreya Gupta (NVIDIA)
Aravind Neelakantan (AWS) |
Thursday, March 19
| Time (PT) | Session | Speakers |
|---|---|---|
| 8:00 AM – 9:45 AM | Ultra Scale Runbook for PyTorch on NVIDIA GPUs for Training and Inference (Hands-on Lab) | Syed Ahmed (NVIDIA), Tianyu Liu & Lu Fang (Meta) |
| 2:00 PM – 2:40 PM | Parameterized CUDA Graph Launch in PyTorch: CUDA Graphs Without the Pain | Daniel Galvez (NVIDIA) |
Pre-NVIDIA GTC: Helion Hackathon — Saturday, March 14
Arriving early or not attending NVIDIA GTC? Kick off the week with the PyTorch Helion Hackathon on Saturday, March 14 in San Francisco!
Join experts from Meta and NVIDIA for a hands-on hackathon focused on kernel authoring with Helion. Whether you’re new to writing GPU kernels or looking to push the boundaries of performance, this is a great opportunity to learn, build, and connect with the community.
👉 Register for the Helion Hackathon
We can’t wait to see you at NVIDIA GTC 2026! Come find us at Booth #338, attend our sessions, and join us for the Helion Hackathon.
See you in San Jose!