PROJECTS
Projects and evidence
Origins, stage and limitations belong together.
First-party research code is separated from adaptations of upstream work. Repositories, documentation and known limitations are the starting point for inspection; public code alone does not validate model performance.
01 / First-party project
KineJing
Evidence-bound experiments for action and world-model research, including a lightweight motion baseline, a visual feature path, and a small trained three-view action-feature predictor.
Research previewFirst-party project↗kine-jepa
Research code exploring compact action-conditioned latent world models for planning and control.
Research codeFirst-party project↗kine-bench
Evaluation and evidence gating for claims about world-model planning.
Research toolingFirst-party project↗kine-datapipe
A research pipeline for turning video and simulation into world-model training data.
Research tooling02 / Open-source adaptation · upstream fork
This category builds on attributed upstream work. Engineering changes do not transfer ownership of upstream model capabilities or validate performance.