Robot learning & manipulation: photo by Ims, CC BY-SA 4.0. Resized; display cropping may apply.
A PROJECT YOU CAN MAKE YOUR OWN
From first questions
to a working demo.
A simulated manipulation demo, benchmark results, and a documented policy implementation.
This is a proposed learning path. The mentor, final project, start date, and weekly schedule will be confirmed before enrollment.
Your four-month journey
MONTH 01Frame the problem+
Define a manipulation task and set up a shared simulation environment.
MONTH 02Build your baseline+
Implement and evaluate a baseline controller or policy.
MONTH 03Experiment & improve+
Experiment with learning methods and examine how your policy fails.
MONTH 04Demonstrate & explain+
Build a reproducible demo and explain your results in a final technical presentation.
What to bring
Comfortable Python, basic probability, and familiarity with machine learning.
Plan for 4–6 hours of independent and team work each week, alongside your one-hour mentor meeting. The pilot is online, using public data and simulation.
Tools you'll work with
Python · PyTorch · Simulation
What you'll take away
- A documented team project and your individual contribution
- A final demo and technical write-up
- Direct feedback and an individual assessment
- Opportunity to request a performance-based reference