The Working Notes: EP3 - How do we prepare data for a World Language Action Model?

Julian Lopez and Konrad Jamrozik from Reka's data platform team walk through what it actually takes to prepare petabytes of video data for world model training, from sourcing and filtering to annotation and deployment. Six pipeline stages. Hundreds of thousands of hours of video. A small team moving fast. Learn more about what we do at Reka Labs: https://reka.ai/labs Claru Marketplace: https://claru.ai/