A glowing pumpkin that spots visitors, hears their questions, describes what it sees and replies spookily, running models on Dragonwing NPU.
Qualcomm IM SDK GStreamer plugins turn Dragonwing IQ-9075 NPU inference into a pipeline element, driving real-time Recamera gimbal tracking
Inspired to develop this local AI agent after failing to remove a yellowjacket nest with my custom bee vacuum due to my lack of knowledge :)
Train LeRobot policies from real demonstrations and deploy on the Dragonwing IQ-9075 NPU for autonomous pick and place task.
This is the 3rd version of what is probably my most popular project: a vintage typewriter (actually a paper terminal) that answers questions
Control a Gazebo simulated TurtleBot 4 using a VLM running on the Dragonwing IQ-9075, translating prompts into autonomous ROS 2 navigation.
Drive the robotic hand to mirror live gestures via NPU-accelerated Qualcomm AI optimized model
Smart device detects shower sound and warns via Home Assistant when the fan is off, preventing steam build‑up and stopping mould growth
This Arduino App Lab project cascades a lightweight TinyML model with a local VLM to bring contextual descriptions of the scene.
Learn how to design, train, and deploy a custom CNN on a DEEPX-accelerated edge device.
A vision ML–enabled garden robot that detects slugs and automatically sprays an organic garlic solution to stop them from reaching plants
I Built a $200 Mini PC That Runs My Personal AI Agent (LattePanda + OpenClaw)
In this project, we take a two-stage OCR pipeline originally built in Python and migrate it to the Arduino UNO Q using Edge Impulse.
An exploration of accelerating the MediaPipe models with AzurEngine
An exploration of accelerating the MediaPipe models with Qualcomm
Autonomous robotic dog leverages ROS 2 and Edge Impulse-trained YOLO-Pro model for real-time ball detection and pursuit on the Arduino Uno Q
An Edge AI-based bottle preparation machine that alleviates the stress of responding quickly when a baby cries in the middle of the night
Develop a simple, low-cost AI-powered wearable with the XIAO nRF52840 Sense that detects hand gestures and controls real-world devices.
Train a domain-specific image to keywords model that is 200x lighter (9MB) and runs with low latency on hardware such as Raspberry Pi 4.
With ROS 2 integration, the YOLO-Pro model from Edge Impulse powers real-time hand tracking on the Rubik Pi 3, accelerated by Qualcomm AI.
Autonomous humanoid robot uses Edge Impulse YOLO-Pro model on Particle Tachyon with Qualcomm AI Accelerator for balls detection and kicking
This project showcases the inference of an Edge Impulse YOLO-Pro model using a 12 TOPS AI Accelerator on the Qualcomm Dragonwing SoC.
A smart helmet for cyclists that recognise audio instructions from the rider and activates LEDs at the back to indicate left or right turns.
An industrial automation system for block sorting, utilizing Computer Vision and Finite State Machines, implemented with Edge Impulse.
