Powering the Gap Between Pixels and Actuators

The Mission

Building Physical AI isn’t just a software challenge — it’s an infrastructure challenge. My work focuses on creating the simulation pipelines and models (RL, DL and more) that allow AI to move from the computer screen into the real world.

Your sponsorship directly impacts the ability to:

Rent High-End GPUs for large-scale Isaac Sim training.
Purchase Sensors & Actuators for “Sim-to-Real” physical validation.
Maintain Open-Source Repositories that lower the barrier for others.

Any Help is appreciated given Increased Cost of Hardware and Cloud Computing

Support via Venmo

If my work has helped you, or if you believe in the future of accessible Physical AI research, consider a direct contribution via Venmo.

📱 Scan to Donate

🚀 Direct Link

Simply click the button below or search for my handle directly on Venmo.

Venmo Handle: @Renan-Barbosa

Donate via Venmo

Every contribution, no matter the size, makes a massive difference in hardware budget and cloud compute time.

Impact Roadmap

The idea is to raise funds until there is enough to do an End-to-End project.

What Your Funds Enable for example:

$6: One hour of cloud GPU on single L40S g6e.8xlarge.
$40: One hour of cloud GPU training time on on 8x L40S g6e.48xlarge.
$400: A Full Kit of the robot SO-ARM101
$500: A Depth Vision Camera IntelRealsense D455

Values are an approximation, donation of hardware and cloud credits are extremely helpful too.

Free Workshops

Currently hold hands-on workshops at Panera Bread every other week: Gathering of Aspiring Technologists (GOAT) Club

Unused Hardware (after project completion) are given away to help students and cloud credits lower the barrier of entry.

What’s Coming Next?

I’m currently working on my trip to GTC 2026

The idea is to make an open Poll plus a discussion forum so we can plan together and vote on which projects to piroritize.

Some projects can be very demanding, for example:

Making a decent VLA project that goes beyond the demos, lets say, replace the backbone VLM model and implement a new skill (within model capability) can take 2-3 weeks and cost around $300 - $500 in cloud computing assuming everything goes according to plan.
Sim2Real projects involve hardware and are naturally more expensive.

Experimental Poll features:

Thank You!

Your support keeps the “Bored Engineer” from actually being bored — and instead, keeps the robots moving.

Back to Projects Check the Blog