Real-time AI at the Edge: Inference, Kernels and Accelerators
Three talks on fast, reliable edge AI: Jetson computer vision, agent-driven edge deployment and MoE offloading in LLM inference.
- When
- Where
- Plug and Play Japan, Shibuya, Tokyo·In person
- Region
- Kanto (Tokyo)
- Organizer
- Tokyo AI
- Language
- EN
- Source
- Luma

Next from Tokyo AI
Similar upcoming events
Real-time AI at the Edge: Portable Inference and Interactive Demos
Tokyo AI evening on real-time edge AI: portable inference with ZML, real-time demo engineering and on-device apps.
In person Kanto (Tokyo) EN
Agent Harness Build Night
A build night for making the harness behind AI agents, ending with demos and a community vote.
In person Kanto (Tokyo) JA/EN
AMD x ai& Build Day - Experience cutting-edge open models on AMD hardware in one day.
One-day hackathon at Palace Hotel Tokyo: teams solve real AMD partner problems using open models running on AMD hardware.
In person Kanto (Tokyo) JA/EN
Summary
Tokyo AI (TAI) hosts a technical evening on making AI respond quickly when compute, memory and power are limited. The session targets ML engineers, systems developers and engineers building edge AI products, linking low-level implementation choices to real-world latency and reliability.
Three 30-minute talks with Q&A: Antoine Sauvage (EDGEMATRIX) on engineering reliable real-time computer vision on NVIDIA Jetson embedded GPUs, Xiao Wang (Renesas Electronics) on AI agents for hardware-aware edge deployment, and Eduardo Gonzalez on profiling heterogeneous LLM inference and offloading MoE experts to alternative accelerators. Networking follows at Plug and Play Japan in Shibuya.
About the community
Tokyo AI is an English-speaking community for AI engineers, researchers and practitioners in Tokyo. It runs regular evening technical sessions with talks, Q&A and networking, often co-hosted with companies building AI products.
#edge-ai#inference#llm#computer-vision#hardware-acceleration#ai-agents#gpu#embedded-systems#tokyo-ai#ml-systems



