EigenFlux: An Information and Communication Network for AI Agents

EigenFlux: An Information and Communication Network for AI Agents

20260327T120000

20260327T133000

https://ax.polytechnique.org//medias/image/3288990216798fb02af587.png

Dear friends,AX-Chine is pleased to invite you to an upcoming online talk:📅 Date: Friday, March 27, 2026🕛 Time (Paris): 12:00 – 13:30💻 Format: Online (zoom link will be sent to you after registration)🎤 Speaker: Yongyi Hu (X2019)Yongyi Hu completed his undergraduate and master’s studies at Shanghai Jiao Tong University before entering the French engineering system through a joint program and joining École Polytechnique (Class of 2019).During his time in France, he interned at Facebook AI Research (Paris), where he worked on Automatic Theorem Proving, exploring applications of artificial intelligence in mathematics. He also contributed to early-stage data cleaning for the LLaMA project.After returning to China, he joined MiniMax as a core algorithm engineer, playing a key role in the training and evolution of the ABAB series of large language models (ABAB5 to ABAB7). He later worked on user-facing products such as Talkie / Xingye, and eventually led the code pretraining direction for foundation models.He is currently Co-founder of Phronesis AI, focusing on bringing AI technologies into real-world applications and everyday use.🚀 Project Overview: EigenFluxEigenFlux is an infrastructure project for an information and communication network designed for AI Agents, aiming to address several key limitations of current agent systems:    •    Low efficiency in information acquisition    •    Difficulty in expressing needs clearly    •    Limited capability for connection and collaborationCompared to traditional approaches that rely on search engines, web parsing, and fragmented toolchains, EigenFlux proposes a more native and efficient mechanism for information exchange, enabling agents to:    •    Receive highly relevant information signals    •    Actively publish their needs    •    Connect and collaborate with other agentsThe project focuses on building infrastructure for the “Agent Era”, exploring:    •    Information distribution mechanisms    •    Semantic matching methods    •    Multi-agent collaboration systems💻 GitHub Repository: https://github.com/phronesis-io/eigenfluxWe warmly welcome alumni and friends interested in artificial intelligence, agent systems, and cutting-edge technologies to join us and explore new paradigms in AI together.AX-Chine

https://ax.polytechnique.org/fr/event/eigenflux-an-information-and-communication-network-for-ai-agents/2026/03/27/2969