What can you build when generative AI moves off the cloud and onto the device?

Join Arm and Meta ExecuTorch Engineers for a two-day, hands-on hackathon exploring what’s possible with AI on the edge using ExecuTorch and Arm-powered hardware.

Meet Your Hardware

You’ll build on the powerful Alif Ensemble E8 development kit, using its AI-accelerated compute capabilities to push the boundaries of what can run directly on an embedded device. Imagine a device that can see and understand its surroundings, recognize what’s happening, and respond intelligently—all without sending data to the cloud. Or speaking to a device and watching it generate images inspired by your conversation, in real time and entirely at the edge.

Working in small teams, you’ll take an idea from concept to working prototype, combining models, sensors, cameras, audio, and other inputs to create intelligent applications that run locally on the device.

Whether you’re an experienced AI/ML developer or already familiar with the broader PyTorch ecosystem and looking to explore edge deployment, this is an opportunity to experiment with on-device AI, collaborate with other developers, and get hands-on support from Arm and Meta experts.

Put your skills to work as you:

  • Learn how ExecuTorch enables efficient AI inference on Arm-powered hardware.
  • Build AI applications designed to run locally on resource-constrained devices.
  • Combine camera, audio, and sensor data to create intelligent, responsive experiences.
  • Experiment with generative AI and other advanced workloads at the edge.
  • Develop creative solutions to real-world—or simply fun—problems using edge AI.

Participants will have access to Alif development kits, technical resources, and guidance from Arm and Meta experts throughout the event.

Apply to Participate

Ready to build with us? Apply to participate by joining the hackathon and submitting a project name and short project overview describing what you’d like to build.

Submitting an application does not guarantee participation in the event.

Space at this in-person event is limited, and participation is subject to availability of the Alif Ensemble E8 development kits. After applications close, Arm will review project proposals and notify selected teams to confirm their place at the hackathon.

Compete, win, and keep building.

Prizes will be awarded to the top three teams, with a dedicated prize for the top all-student team. So whether you’re already working in the industry or looking to put your skills to the test as a student, you’ll have an opportunity to showcase what you can build with AI at the edge. 

And the opportunity doesn’t end when the hackathon does. Winning participants will be invited to become ExecuTorch Community Ambassadors, opening the door to continued involvement with the community and the opportunity to participate in Arm's ExecuTorch pilot program. Those accepted will gain early access to emerging projects and technologies around the Arm ExecuTorch backend—and help shape what comes next for AI at the edge.

Requirements

Each team should be prepared to provide the following as part of their final submission:

  • Live Demonstration: A working demonstration of the project running on the Alif E8 Dev Kit.
  • Public Source Code Repository: A publicly accessible repository containing the source code for the project.
  • Fixed Release Version: A commit hash or release tag identifying the exact version of the code used for the judged demonstration.
  • Build and Run Instructions: A README with clear instructions for building, deploying, and running the project.
  • Application Architecture: A brief description of the application architecture and its major components.
  • Workload Placement: Identification of where key models and workloads execute within the system.
  • Performance Measurements: Basic performance measurements, such as latency, memory usage, or other relevant metrics, where available.

Submission ownership

Submisison must be the original work of the Entrant, be solely owned by the Entrant, and not violate the IP rights of any other person or entity.

Language Requirements

All Submission materials must be in English or, if not in English, the Entrant must provide an English translation of the demonstration video, text description, and testing instructions as well as all other materials submitted. 

Intellectual Property 

Your Submission must: (a) be your (or your Team, or Organization’s) original work product; (b) be solely owned by you, your Team, your Organization with no other person or entity having any right or interest in it; and (c) not violate the intellectual property rights or other rights including but not limited to copyright, trademark, patent, contract, and/or privacy rights, of any other person or entity. An Entrant may contract with a third party for technical assistance to create the Submission provided the Submission components are solely the Entrant’s work product and the result of the Entrant’s ideas and creativity, and the Entrant owns all rights to them. An Entrant may submit a Submission that includes the use of open source software or hardware, provided the Entrant complies with applicable open source licenses and, as part of the Submission, creates software that enhances and builds upon the features and functionality included in the underlying open source product. By entering the Hackathon, you represent, warrant, and agree that your Submission meets these requirements.

Financial or Preferential Support 

A Project must not have been developed, or derived from a Project developed, with financial or preferential support from the Sponsor or Administrator. Such Projects include, but are not limited to, those that received funding or investment for their development, were developed under contract, or received a commercial license, from the Sponsor or Administrator any time prior to the end of Hackathon Submission Period. The Sponsor, at their sole discretion, may disqualify a Project, if awarding a prize to the Project would create a real or apparent conflict of interest.

 

Hackathon Sponsors

Prizes

1 non-cash prize
Ray·Ban Meta AI Glasses
4 winners

All members from each of the winning teams will receive a pair of Ray·Ban Meta AI Glasses

Devpost Achievements

Submitting to this hackathon could earn you:

Judges

Darrell Malone

Darrell Malone
Arm

TBD

TBD
Arm

TBD

TBD
Meta

TBD

TBD
Meta

TBD

TBD

Judging Criteria

  • Functionality and Reliability - 20 points
    Does the application work reliably on the Alif E8? Is the demo repeatable and stable?
  • Performance and Efficiency - 20 points
    End-to-end latency, FPS/throughput, responsiveness, memory utilisation and power/energy where measurable. Evidence of optimisation is encouraged.
  • Effective use of the Hardware - 15 points
    Appropriate use of the available CPU, NPU, memory and hardware peripherals. Reward good architectural decisions rather than simply maximising NPU usage.
  • Multimodal and Sensor Integration - 15 points
    Meaningful use of cameras, microphones and/or sensors. Reward quality of integration and sensor fusion rather than simply the number of inputs used.
  • Innovation and Use-Case Value - 10 points
    Originality, usefulness, relevance to edge AI, and how effectively the application demonstrates the possibilities of the platform.
  • Technical Ambition and Integration - 10 points
    How far the team goes beyond the starter examples: custom models, heterogeneous compute, multiple models, sensor fusion or sophisticated pipelines.
  • Engineering Quality and Demo - 10 points
    Code quality, reproducibility, documentation, architectural explanation and quality of the final demonstration.

Questions? Email the hackathon manager

Invite others to compete

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.