I build this Open-Source AI Robot a custom body and app!

I gave my Reachy Mini a makeover — then made her actually useful How to build a custom AI app for the Reachy Mini robot. We're using TensorFlow and PyCharm to build real-time object tracking on this open-source platform from Pollen Robotics. Reachy Mini is one of the most interesting robots out right now: built by Pollen Robotics in collaboration with Hugging Face and Seeed Studio, fully open source (including the body), and genuinely AI-first in a way most robots aren't. I got my hands on a beta developer unit and immediately did the most important thing: dressed her up. After printing a custom pink outfit with my name on it (thanks to the Elegoo Centauri Carbon 2 and its multicolor printing), we actually get into what makes her interesting as a platform: the app ecosystem, the camera-based demos, and then building a custom TensorFlow object detection app in PyCharm and deploying it live. Head tracking, rock paper scissors, real-time object tracking: all running on a robot the size of a desk lamp. She's not a manipulation robot. She can't move around or grab things. But as a demo for where AI-first robotics is going, she's a surprisingly good starting point. Full code breakdown (TensorFlow + PyCharm setup) → ‪@PyCharmIDE‬    • Build a TensorFlow Object Detection App | ...   Reachy Mini app on Hugging Face → https://huggingface.co/reachy-mini Pollen Robotics Discord (where I found the body files) →   / discord   Chapters 00:00:00 Intro — meet Reachy Mini 00:00:50 The body is open source (so obviously I had to print stuff) 00:01:07 Multicolor printing with the Elegoo Centauri Carbon 2 00:02:31 Plugging her in — hardware overview 00:04:49 Running AI locally vs. in the cloud 00:05:33 The app store — what people have built 00:06:00 Rock paper scissors with computer vision 00:06:08 Head tracking demo 00:07:11 Building a TensorFlow object detection app in PyCharm 00:07:39 Live object detection — it works 00:08:26 What Reachy Mini actually is (and isn't) 00:09:58 Wrapping up + what's next Sponsor Callouts This video is sponsored by PyCharm from JetBrains. I’ve used their awesome IDE for Python Development to build out the integrated computer vision app (with TensorFlow!). The second video will go live on their channel directly, where I walk you through the notebook I wrote and how I built out the Reachy mini app using the IDE features and some help from Claude. Elegoo sponsored this video through a gifted 3D printer: the Centauri Carbon 2. They sent me the 3D printer to test out, and for this project, and I was particularly happy with the color switching capabilities! This made the customization with my brand possible for the Reachy outfits! You can check it out via my (affiliate) link here: https://tidd.ly/41Zf06V #elegoo ‪@ElegooOfficial‬ During this month, Rerun is also supporting Back To Engineering. Follow iulia on: Patreon for updates, chat, extra content: patreon.com/c/BacktoEngineering Instagram for behind the scenes: instagram.com/backtoengineering GitHub for the project code: https://github.com/iuliaferoli X for yapping: x.com/iuliaferoli For business enquiries (sponsorships, collabs, etc) reach out at: [email protected] #robotics #opensource Thank you to my sponsors for making this video possible!