This hands-on workshop guides participants through building a multiplayer Battleship game powered by computer vision AI on NVIDIA Jetson devices running Avocado OS. Attendees will create a complete system that uses cameras to detect physical game pieces, processes vision data through machine learning models, and synchronizes game state across multiple players using Elixir and Phoenix.
The workshop demonstrates how Elixir’s fault-tolerant architecture makes it ideal for embedded AI applications, allowing developers to create resilient systems that can handle hardware constraints, network disruptions, and component failures. Participants will learn practical techniques for deploying computer vision inference on edge devices while maintaining high availability through supervision trees and distributed state management.
Beyond the engaging Battleship implementation, attendees will gain valuable architectural patterns applicable to commercial embedded products across industrial, IoT, and consumer applications. This workshop bridges the gap between theoretical concepts and practical implementation, delivering both immediate satisfaction through gameplay and long-term value through production-ready embedded system design principles.
Workshop prerequisites: This workshop requires additional hardware, which participants will need to purchase separately (estimated cost: approx. $100). Details on how to obtain the hardware will be provided before the event.
DURATION
3 hours
Workshop objectives:
- Equip developers with practical skills for building AI-powered devices using Elixir and Avocado OS
- Demonstrate how to implement and optimize computer vision inference on NVIDIA Jetson hardware
- Showcase Elixir’s strengths in creating resilient embedded AI applications
- Provide a complete pipeline from camera input to AI processing using the Avocado OS ecosystem
- Create engaging, real-world examples through multiplayer interactive systems
- The ultimate goal is enabling participants to leverage Elixir’s concurrency model and fault-tolerance for next-generation embedded AI products