Open Source Models & Infra
AI-Ready Data Production and Dynamic Data Training for Open-Source Models
Date, time, and room will be added once confirmed.
Talk overview
The talk discusses the paradigm shift toward high-quality data dependency for model performance, introducing a unified data flywheel with DataFlow L0–L7 open-source infrastructure. It explores agentic execution through agents, skills, and harnesses, along with a closed-loop evolution driven by evaluation. The session also covers engineering practices for implementing a data-centric AI open ecosystem.
