Research
Run a reproducible manipulation experiment
Build a controlled arm task, establish a baseline and report failure modes before adding complexity.
A ROLE. A TASK. A PATH FORWARD.
Research & robot learning
Design reproducible experiments across manipulation, imitation learning, embodied models, locomotion and perception.
Start at your current level. Move forward when you can demonstrate the milestone. These editorial paths describe project depth; they do not award a professional qualification.
Choose your role and goal
Explore the full supporting collections
The stages select a focused set of starting resources. Browse these directory categories for alternatives and deeper specialization.
Research scenarios
6 paths · 3 stages eachResearch
Collect demonstrations and train an imitation policy
Develop a traceable teleoperation, dataset, training and held-out evaluation workflow.
Research
Evaluate vision-language-action models responsibly
Compare embodied models under matched task, dataset and execution assumptions.
Research
Study locomotion and humanoid control
Move from a dynamics model to controlled learning experiments and supervised transfer planning.
Research
Build a perception dataset and evaluation study
Organize sensing, calibration, labels and held-out evaluation for a bounded robotics perception task.
Research
Plan a shared robotics research lab
Choose platforms around experiments, operating procedures and reproducible onboarding.