GMI x NVIDIA Night @CVPR

​At CVPR, the frontier is shifting —
from models to systems that perceive, reason, and act.

​Join NVIDIA and GMI Cloud for an evening exploring how multimodal AI is evolving into real-world applications —
from VLMs to agentic vision and beyond.

​What we’re focusing on

• ​real outputs, not just research

• ​systems, not isolated models

• ​deployment, not just benchmarks

​From paper → output

​Using GMI Studio, your paper can be turned into:

• ​a visual poster

• ​ready-to-share social media content (text + image)

​Selected works will be:

• ​showcased live during the event

• ​amplified via GMI’s X and LinkedIn

​👉 Free for CVPR Researchers (Limited Time) Turn your research into shareable content in minutes:
[Insert Link — Paper → Poster & Social Generator]
1. Research Paper to Poster
2. Research Paper to Social Post

​Join our Discord to claim your free credits and share feedback with the team

​Please upload your outputs on the form to be considered a demo slot!

​Inside the night

1. ​6:30 PM - Inside GMI Model Library: One API, Every Modality

   • ​A live walkthrough of the GMI Model Library and how we empower multimodal generation through a single unified API layer

2. ​6:45 PM - GMI Studio Demo

3. ​Researcher Lightning Showcases
   Selected CVPR authors present: their work and generated poster + social outputs

   • ​7:05 PM - Allen Tu · University of Maryland (UMIACS) SpeeDe3DGS & SplatSuRe — CVPR 2026 Oral

   • ​7:10 PM - Minjae Jeong · POSTECH Mark4D: Temporally-Consistent Watermarking for 4D Gaussian Splatting — CVPR 2026 Oral

   • ​7:15 PM - Federico Spurio · University of Bonn Video Panels for Long Video Understanding — CVPR 2026 Oral

   • ​7:20 PM - Andres Villa · KAUST CURE: Curriculum-guided Multi-task Training for Reliable Anatomy Grounded Report Generation — CVPR 2026 Oral

   • ​7:25 PM - Yifan Jiang · USC / Amazon VideoP2R: Video Understanding from Perception to Reasoning

   • ​7:30 PM - Chengqi Duan - CodePlot-CoT: Mathematical Visual Reasoning by Thinking with Code-Driven Images

​Why Attend

• ​Understand where vision AI is heading:
  → multimodal → agentic → real-world systems

• ​Happy hour with gourmet food & drinks

• ​Connect with:

  • ​top CVPR researchers

  • ​AI builders and startups

  • ​infrastructure and platform leaders

​Who Should Join

• ​Researchers in vision, multimodal AI, robotics, or generative models

• ​Engineers building multimodal or agentic systems

• ​Startups in image, video, or embodied AI

• ​Teams exploring real-world deployment of AI systems

​Hosts

​GMI Cloud — one of seven NVIDIA Reference Platform Cloud Partners globally, powering the production layer of multimodal AI

​NVIDIA — shaping the future of physical AI, agentic vision, and world models

​See you in Denver.