About the challenge

We’re bringing together 30 builders, hackers, and robotics engineers to tackle a unique, high-energy theme: Cooking with Robot Arms! Whether it’s flipping pancakes, sorting ingredients, pouring liquids, or executing delicate culinary assembly, Revolute is all about bridging the gap between open hardware, edge compute, and foundation vision-language-action (VLA) models.

Our core mission isn't just competition—it’s bringing together and growing the Boston robotics and physical AI maker community. Expect great food, awesome gear, friendly mentors, and a weekend full of creative engineering!

Get started

  • Provided Hardware per Team (4–5 people): Each team gets access to a physical kit including one SO-100 follower arm, one leader control arm, two USB cameras, and an NVIDIA Jetson edge compute system.

  • The Goal: Take your arm setup from zero to executing a physical task—whether through teleoperation dataset collection, fine-tuning policies (like LeRobot or NVIDIA Isaac GR00T), or building custom vision-control loops.

  • Supported Backers & Partners: Backed by Seeed Studio, Fab Foundation, MITERS, and Boston Robot Hackers

Requirements

What to Build

  1.  

Build a functional software-hardware pipeline where your robotic arm setup executes a physical manipulation task (bonus points for leaning into the culinary/cooking theme!). This can range from:

  • Autonomous Manipulation: Teleoperated or policy-driven pick-and-place, pouring, chopping, or sorting.

  • VLA & Policy Fine-Tuning: Training models (e.g., LeRobot pipelines, Isaac GR00T, or custom neural nets) on your teleoperation data and running inference on your Jetson edge unit.

  • Open Hardware Extensions: Custom 3D printed grippers, utensil attachments, or multi-camera mounts integrated with the arm.

 

What to Submit

  1.  

1. Project Overview & One-Liner
2. Demo Video (2–3 Minutes) System Breakdown: Briefly narrate or overlay text showing:

-Teleoperation / Dataset collection phase.

- Training / Fine-tuning approach (e.g., LeRobot, GR00T, ACT, Diffusion Policy).

- Real-time inference running on the Jetson unit.

  1. System Architecture & Story

Break down your build step-by-step using photos, diagrams, or code snippets:

  • Things Used / Bill of Materials: List all hardware (SO-100 Leader/Follower, Jetson model, cameras, custom 3D-printed end effectors) and software frameworks used.

  • Data Collection & Teleoperation: How did you record your demonstrations? Mention total episodes recorded, camera angles, and dataset preparation.

  • Model Training & Fine-Tuning: Explain your training pipeline, policy selection (e.g., Hugging Face LeRobot, NVIDIA Isaac GR00T / Isaac Sim, Pi0), and hyperparameters.

  • Edge Deployment: How did you deploy and optimize the model for real-time inference on the Jetson edge hardware?

  • Challenges & Lessons: What physical, driver, or sim-to-real roadblocks did you encounter, and how did you debug them?  

 
4. System Architecture & Story 

  • our GitHub repo link must be public and include:

    • Codebase: All scripts for teleop recording, policy evaluation, model inference, and server configs.

    • Custom Assets: CAD files (.STL / .STEP) if you designed custom utensil holders, grippers, or mounts.

    • Step-by-Step README.md: Clear terminal commands so anyone can clone your repository and reproduce your setup on their own SO-100 arm and/or Jetson setup.

 

Prizes

3 non-cash prizes
1
1 winner

$200 per person, can be used toward any reBot Arm Kit(RS Kit\DM Kit\Leader Arm) product: https://www.seeedstudio.com/reBot-Arm-B601-RS-Disassembly-Kit-Version-with-Power-Supply-Bundle.html

2
1 winner

Seeedstudios Hugging face products（SO-ARM101/Lekiwi/Amazing Hand）: $160 per person, can be used on any of the following product pages:
SO-ARM101:https://www.seeedstudio.com/SO-ARM101-Pro-Unassembled-Bundle.html
Lekiwi:https://www.seeedstudio.com/LeKiwi-Full-Kit-12V-Verision.html
AmazingHand:https://www.seeedstudio.com/Amazing-Hand-Right-Hand-The-Open-Source-Robotic-Hand-Developer-Kit.html

3
1 winner

$80 per person, can be used on any of the following products [ ]
fashionstar mini robot
https://www.seeedstudio.com/Fashionstar-Atom-X-17-DOF-Open-Source-Robot-Kit-p-6712.html
stakeforce mini wheeled-robot
https://www.seeedstudio.com/StackForce-Mini-Wheeled-Legged-Robot-p-6710.html

Judges

Sayam Dhingra

Sayam Dhingra
AutoDesk

Judging Criteria

  • Documentation & Execution (40 Points)
    Live demo, GitHub code, presentation deck, and Devpost video.
  • Technical Difficulty (30 pts):
    Ambition, accuracy, and reliability.
  • Product & Taste (30 pts):
    Creativity, edibility, and overall chef approval.

Questions? Email the hackathon manager

Tell your friends