Uploaded October 2026 | Updated November 2026, 4 hours ago
A six-wheel rover with a Raspberry Pi 5, a camera, LiDAR, and a brain running on two GPUs in Akamai Cloud. Say "hey rover, find the green ball" and it looks around, tells you where it is, follows it, and takes a picture on command.
In this video I walk through the build and then let it loose. The rover is a Waveshare UGV Rover with a Raspberry Pi 5. Camera and mic traffic hits a Caddy gateway and routes to a single Linode with two RTX 4000 GPUs, 20GB of VRAM each. YOLO handles object detection, Whisper handles voice transcription, Qwen is the agent that reasons and makes tool calls, and text to speech talks back. Every model and its context fits on those two GPUs. You'll see live object detection and transcription in the browser, the rover tracking and following a blue ball (and finding it again after I hide it), voice commands to find and follow the green ball, a tool call that takes a photo, and a joke the text to speech almost gets right. Code is linked below.
### 🔗 Useful Links & Resources:
* Akamai Developers on GitHub: github.com/akamai-developers
* Akamai Developer Hub: developers.akamai.com
* $100 Akamai Cloud Credit: login.linode.com/signup?promo=akm-lp-akmtube-100-111325-T853
### 📲 Follow Us & Join Our Community
🆇 X: https://x.com/akamaidevs?s=20
💼 LinkedIn: linkedin.com/showcase/akamai-dev
👾 Discord: discord.gg/5MJegzS8Y
### ⏱️ Timestamps:
0:00 Waveshare UGV Rover With Raspberry Pi 5, Camera, and LiDAR
0:40 How It Works: Caddy Gateway, YOLO, Whisper, Qwen, and Text to Speech
1:32 Architecture: One Linode, Two RTX 4000 GPUs on Akamai Cloud
2:10 Live Demo: Object Detection and Voice Transcription in the Browser
3:04 Find Object: Tracking and Following the Blue Ball
4:32 Voice Agent: Hey Rover, Find the Green Ball
5:38 Tool Calls: Follow the Ball and Take a Picture
6:11 Hey Rover, Tell Me a Joke
7:04 Get the Code
A six-wheel rover with a Raspberry Pi 5, a camera, LiDAR, and a brain running on two GPUs in Akamai Cloud. Say "hey rover, find the green ball" and it looks around, tells you where it is, follows it, and takes a picture on command.
In this video I walk through the build and then let it loose. The rover is a Waveshare UGV Rover with a Raspberry Pi 5. Camera and mic traffic hits a Caddy gateway and routes to a single Linode with two RTX 4000 GPUs, 20GB of VRAM each. YOLO handles object detection, Whisper handles voice transcription, Qwen is the agent that reasons and makes tool calls, and text to speech talks back. Every model and its context fits on those two GPUs. You'll see live object detection and transcription in the browser, the rover tracking and following a blue ball (and finding it again after I hide it), voice commands to find and follow the green ball, a tool call that takes a photo, and a joke the text to speech almost gets right. Code is linked below.
### 🔗 Useful Links & Resources:
* Akamai Developers on GitHub: github.com/akamai-developers
* Akamai Developer Hub: developers.akamai.com
* $100 Akamai Cloud Credit: login.linode.com/signup?promo=akm-lp-akmtube-100-111325-T853
### 📲 Follow Us & Join Our Community
🆇 X: https://x.com/akamaidevs?s=20
💼 LinkedIn: linkedin.com/showcase/akamai-dev
👾 Discord: discord.gg/5MJegzS8Y
### ⏱️ Timestamps:
0:00 Waveshare UGV Rover With Raspberry Pi 5, Camera, and LiDAR
0:40 How It Works: Caddy Gateway, YOLO, Whisper, Qwen, and Text to Speech
1:32 Architecture: One Linode, Two RTX 4000 GPUs on Akamai Cloud
2:10 Live Demo: Object Detection and Voice Transcription in the Browser
3:04 Find Object: Tracking and Following the Blue Ball
4:32 Voice Agent: Hey Rover, Find the Green Ball
5:38 Tool Calls: Follow the Ball and Take a Picture
6:11 Hey Rover, Tell Me a Joke
7:04 Get the Code










