Hack the edge.

A 2-day, in-person hackathon in Tokyo where builders come together to push Liquid AI's Liquid Foundation Models to their limits — working with LFM2.5, LFM2.5-VL, and the upcoming LFM2.5-Audio-JP to build Japan-market applications on AMD Ryzen AI PCs.

Spend four weeks pushing the boundaries of Earth observation. Build real-world applications powered by Liquid AI Vision-Language Models (LVMs) using live satellite imagery and space data.

​Build real-world applications powered by Liquid Foundation Models (LFMs) running on AMD NUCs with ROCm™ acceleration -- compact systems engineered for high-performance AI. The weekend also featured fireside conversations, the first led by Pau Labarta Bajo (Developer Relations, Liquid AI) and Ramine Raone (CVP of AI, AMD); followed by Bryan Madden (Global Head of Marketing, AMD) with Ramin Hasani (CEO & Co-Founder, Liquid AI). Together, they explored the momentum behind on-device intelligence, the shift toward hybrid compute, the growing capability of small optimized models, and how open ecosystems empower developers to move faster.

Hack #03 lit up Tokyo with Liquid AI Nanos. Our first in-person hackathon with Weights & Biases and Lambda AI brought builders together in Japan's capital to push Liquid AI's nanos to their limits. Builders created apps that save lives, translate manga, and detect dementia. Fast. Private. On-device. No cloud, no limits.

Hack #02 gave builders their chance to hack with vision. LFM2-VL brought real sight to the edge, and LEAP made it simple to deploy directly into iOS and Android apps. Fast. Private. On-device. No cloud, no lag, no compromises.

Hack #01 challenged builders to rethink control. LFM2 and LEAP made it simple to deploy small language models directly into iOS and Android apps — fast, private, and on-device.