Where hardware meets intelligence. Where code meets the physical world.

Welcome to RoboHack — a hackathon for builders, tinkerers, and AI engineers who want to do more than push pixels. Over the course of this event, you'll either bring a robot to life from a pile of parts, or teach an existing robot to think, see, and act. Whichever path you choose, you're walking out with something that moves.

Pick your track. Form your team. Make a robot do something it couldn't do yesterday.

Build it. Wire it. Make it move.

Your team starts with a LeRobot kit and an empty workbench. By the end, you'll have a fully assembled, calibrated robotic arm executing real-world tasks — and ideally, doing things its designers never intended.

This track is for the people who light up at the smell of a soldering iron and the whir of a 3D printer. You'll be printing parts, assembling mechanical linkages, wiring up electronics, flashing firmware, and writing the low-level control code that turns voltage into motion.

What you'll do:

• Assemble and 3D print a robotic arm from a LeRobot kit
• Calibrate joints, sensors, and actuators from scratch
• Develop control systems for real manipulation tasks
• Push the hardware past its stock capabilities — custom end-effectors, modified kinematics, novel sensing, whatever you can dream up

Skills: Mechanical Design · Electronics · Low-Level Programming · 3D Printing

Perfect for: Hardware hackers, robotics students, mechatronics nerds, and anyone who's ever wanted an excuse to learn inverse kinematics.

Give a robot a brain. Then give it a job.

Your team works with humanoid and quadruped platforms to develop the intelligence that makes them genuinely useful. The hardware is ready — your job is to make it think. Autonomy, perception, decision-making, real-time interaction with the messy physical world.

This is where your favorite ML model finally escapes the Jupyter notebook. Train it, deploy it, and watch it succeed (or fail spectacularly) on actual hardware. Both outcomes make for great demo footage.

What you'll do:

• Program autonomous behaviors and decision-making
• Build navigation, interaction, or task-execution pipelines
• Integrate vision, language, or policy models into real-world control loops
• Get a humanoid or quadruped to do something genuinely impressive on demo day

Skills: Machine Learning · Computer Vision · Autonomy · ROS

Perfect for: AI/ML engineers who are tired of benchmarks and want their models to live in the real world.

• Ambition — pick something hard and go for it
• Working demos — a robot that actually does the thing beats a deck that describes the thing
• Creativity — surprise us
• Documentation — show your process, your failures, and your wins

Whether you're a robotics PhD, a self-taught maker, an ML engineer curious about embodiment, or a CS student who's never touched a screwdriver — there's a place for you here. Mixed teams across both tracks are encouraged, and often the most fun.

• Access to LeRobot kits, humanoid and quadruped platforms
• Mentorship from robotics and AI engineers
• Workspace, tools, and a 3D printer farm
• The satisfaction of making metal and silicon do your bidding

Bring your laptop. Bring your ideas. We'll bring the robots.

Let's build.

• Technical execution
  A maximum of 25 points will be awarded for the technical execution of your project.
• Performance / functionality
  A maximum 25 points will be awarded for the performance/functionality of your project.
• Creativity / originality
  A maximum 20 points will be awarded for the creativity and originality of your project.
• Prototype quality
  A maximum of 15 points will be awarded for the quality of your prototype.
• Presentation / pitch clarity
  A maximum of 15 points will be awarded for the quality of your presentation and pitch.