Code, Hardware, and a Snowstorm: how 1,000 developers built the future of edge AI
Sign up for Developer monthly newsletter
Join thousands of developers around the globe who receive latest news and updates from our monthly curated newsletter.
Sign up
Come for support, stay for the community
Get support from experts, connect with like-minded developers, and access exclusive virtual events.
Join Developer DiscordWhile a major snowstorm blanketed the MIT campus, Qualcomm Technologies was inside fueling a collaborative "storm" of our own at MIT Reality Hack 2026. Nearly 1,000 innovators from around the globe gathered for this premier experiential hackathon to explore the theme "Dreaming Machines."
Over five days, participants pushed the boundaries of on-device AI and spatial computing, turning conceptual ideas into functional prototypes.
To support these builders, our Developer Marketing team actively prepared participants by conducting a pre-hack online information sharing and Q&A session, followed by four intensive in-person workshops during the event.
To further empower the next generation of tech talent, we provided over $5,000 in cash prizes across four specialized categories: Best XR Experience, Best AR Utility, Best Use of Edge AI, and Best Use of Arduino App Lab.
In addition to these awards, standout developers had the opportunity to explore internship roles at Qualcomm, connecting their hackathon success directly to potential career paths.
Beyond the hack: expanding our innovation frontiers at MIT
This year, we expanded our footprint by launching a new initiative alongside our core hackathon, leading critical discussions at the Experiential Innovation Conference at MIT.
This prestigious event brought together world-class researchers, creators, and executives to explore how AI-driven immersive technologies and spatial computing are reshaping human perception and interaction.
A major highlight was the introduction of the Arduino UNO Q board - a collaboration that brings Qualcomm Technologies’ high-performance edge AI to a radically accessible, open-source format. This lowers the barrier for developers, from students to professional engineers, to prototype sophisticated, low-latency applications.
To anchor these discussions, our participation featured two distinct keynote sessions led by Ziad Asghar and Marcello Majonchi, each offering a unique perspective on the future of intelligent, human-centric technology.
AI + XR: transforming human-machine interaction across devices
In his keynote, Ziad Asghar, SVP & GM of XR, Wearables & Personal AI, Qualcomm Technologies, shared how AI and spatial computing are merging to transform the personal devices we use every day, including smartphones, wearables, and XR headsets.
He explained the industry move toward "hybrid AI"- where processing happens both in the cloud and directly on the device and highlighted why gathering data at the "edge" is becoming so critical.
By leveraging Qualcomm Technologies’ latest breakthroughs in efficient computing and connectivity, Ziad showcased how we are creating more natural, intelligent interfaces. These advancements allow devices to work together seamlessly, empowering developers to build highly personalized and immersive experiences that feel more intuitive than ever before.
Human-centric innovation: building the future the “Arduino Way”
Marcello Majonchi, Senior Director, Product Management Arduino for Qualcomm Europe, delivered a compelling keynote on making advanced technology radically accessible through a design philosophy rooted in simplicity and openness. Tracing Arduino’s journey from its origins in Italy to a global ecosystem of 33 million creators, he shared how the company empowers users to build meaningful solutions without requiring deep technical expertise.
The session served as the formal introduction of Arduino UNO Q, powered by Qualcomm technology.
By combining democratized hardware with modular tools like the Arduino App Lab, Marcello demonstrated how they are lowering barriers to electronics, IoT, and edge AI to foster a more inclusive and practical environment for the global developer community.
The Core of Our Presence: The Hack
While strategic discussions took place in the conference rooms, real breakthroughs happened on the hackathon floor. This year’s focus is centered on the fusion of hardware and software across two primary tracks: Immersive XR and the Edge AI Hardware Hack.
Tools of the trade
To help hackers turn the "Dreaming Machines" theme into reality, we provided a versatile toolkit of hardware and software development tools designed for high-performance, on-device intelligence:
- Arduino UNO Q & Arduino Modulino nodes: As the centerpiece of the hardware track, the Arduino UNOQ allowed participants to execute resource-efficient machine learning workloads locally.
This was complemented by Modulino which are plug-and-play hardware nodes that enabled hackers to quickly add sensors and actuators to their projects. By pairing this hardware with the Arduino App Lab IDE, developers gained access to real-time telemetry and a streamlined workspace to validate sensor data and AI inference without a cloud connection. - Edge Impulse : To bring intelligence to the hardware, builders utilized Edge Impulse. This platform served as an end-to-end environment for training and deploying custom models.
Teams used it to build image classification and object detection (FOMO) models, optimizing them specifically to run natively on the Arduino UNO Q with minimal latency. - Next-Gen Spatial Hardware (XR Track): Participants explored the new frontier of spatial computing, leveraging advanced hardware to bridge the physical and digital worlds.
- Samsung Galaxy XR: Powered by the Snapdragon XR2+ Gen 2 platform and the Android XR stack, this hardware enabled high-fidelity immersion. By integrating the Gemini API for Unity, developers fused camera-based world-tracking with responsive spatial interactions, leveraging voice and hand-gestures to create real-time environments that react to the user’s surroundings.
- RayNeo X3 Pro for AR : Powered by the Snapdragon AR1 Gen 1 platform and the RayNeo OpenXR ARDK, hackers focused on gaze-based interaction and anchored spatial overlays to build context-aware tools for smart glasses.
Winning Projects
The level of innovation was staggering. Here are the projects that rose to the top:
- Best Use of Edge AI (Hardware Hack): SoundSense
The SoundSense team developed an on-device inference system to provide the hearing impaired with haptic and visual alerts for environmental triggers like knocks or alarms. The hardware stack centers on the Arduino UNO Q, running a pruned and quantized Whisper Lite model that handles all audio feature extraction and inference locally.
By leveraging Edge Impulse-compatible pipelines for preprocessing and model optimization, the developers successfully eliminated cloud dependency and minimized latency. The project won this category for its technical execution of high-performance audio classification on the Arduino UNO Q, demonstrating a sophisticated integration of edge hardware and optimized software workflows.
- Best Use of Arduino App Lab: Oops All Motion
The team developed a high-speed, on-device motion coaching system for physical rehab and wellness using the Samsung Galaxy XR headset and the Arduino UNO Q.
By capturing motion sequences via the headset’s onboard cameras and training pose-analysis models in Edge Impulse, they built a low-latency pipeline for real-time form correction. The workflow involved optimizing these models for on-device execution and deploying them via Arduino App Lab as modular building blocks to ensure a private, portable user experience. This project won for its technical integration of complex computer vision models into a wearable form factor, enabling high-speed edge inference without cloud dependency.
- Best AR Utility: Forget Me Not
The team built a cognitive aid for Alzheimer's patients using the RayNeo X3 Pro to provide non-intrusive memory assistance. The technical stack integrates privacy-preserving facial recognition and local feature embedding with a conversational agent to help users identify loved ones and navigate safely.
Leveraging the ARDK, the developers implemented a gaze-anchored UI for persistent data overlays and used custom debugging interfaces to optimize anchor stability and identity confidence. The project won this category for its sophisticated use of AR utility - blending complex computer vision with a low-latency, on-device conversational companion designed to preserve user independence.
- Best XR Experience: Reboot: Anomaly Holdout
The team developed a sci-fi survival game for the Samsung Galaxy XR that bridges the digital and physical worlds through a custom-built control panel. Using a microcontroller to interface 3D-printed buttons, dials, and levers, the developers mapped physical inputs to in-game logic via Unity’s Gemini API.
The technical execution relied on mixed-reality passthrough to layer high-pressure game states over the physical hardware, validated by a custom input-visualizer app to ensure signal synchronization. The project won for its successful integration of hardware with immersive XR, delivering a tactile and low-latency hybrid gameplay experience.
The "Dreaming Machines" of MIT Reality Hack 2026 are just the beginning. The creativity we saw, even in the midst of a snowstorm, proves that the next generation of spatial computing and Edge AI is already here.
To see even more incredible ways developers used Qualcomm technology during the event, explore the full project gallery on the MIT Reality Hack Devpost.
Ready to create your own "Dreaming Machine"?
We want to see what you can build. Whether you are interested in immersive gaming, health tech, or on-device AI, here is how you can get started today:
Get Hands-on with the Hardware:
· Get started with Arduino Uno Q: Build with the first-ever Arduino board powered by Qualcomm Dragonwing QRB2210.
· Get started with Samsung Galaxy XR: Develop for the next generation of Android XR.
· Get started with RayNeo X3 Pro: Create contextual AR experiences on the Snapdragon AR1 platform.
Continue the journey with us:
The breakthroughs we saw at MIT were the result of a vibrant, supportive ecosystem of engineers and creators. We want to extend that same support to you:
Discord: Join our developer discord community to chat with Qualcomm Technologies engineers and collaborate with fellow builders.
Events: Our teams are hitting the road! Check our Developer Events page to see if we’re coming to a hackathon or conference near you.
