Fly Brain & Tiny Machines: A BioCompute Hackathon by Neuraffica
A humble fruit fly can fly, land and dodge a swatter on a brain smaller than a poppy seed. Its wiring has now been mapped synapse by synapse and released as open data. For one day, we'll put that wiring to work. And have some fun doing it!
The challenge
Use the fly’s connectome to make a machine see, move or decide better at the edge, and measure it. Small networks run faster on less compute. Treat the fly’s wiring as a blueprint for AI that is capable and tiny.
The rules
The brain has to matter. Fly connectivity must shape your model, not just appear on a slide.
Compare against a standard baseline and show the numbers.
Run it on a device, ours or yours.
Ideas to get you started
Image classification or a small vision-language model built on the fly’s visual circuits
Looming detection on a micro-drone
A robot dog gait driven by locomotion circuits
An RL policy shaped like the connectome
Something nobody has tried
What we provide
The fly connectome and a starter kit that turns circuits into runnable models
Hardware: micro-drones with onboard AI chips, robot dogs and small robot cars, plus a simulator and open datasets
Breakfast, lunch, snacks and coffee all day
Prizes: the winning team picks hardware to take home
Judges from the deep-tech and investor community
Who should apply
ML and RL engineers, computer vision people, robotics and embedded hackers, and neuroscientists who code. Come solo or with a team, or form one on the day. You’re welcome to bring your own robot or dataset; just make sure it works.
Schedule
9:00 Doors open, breakfast, and a conversation with the Neuraffica founders
10:00 Rules, team setup, device checkout
11:00 Hacking begins
16:00 Submissions due
17:00 Finalist pitches
18:30 Winners announced
Until 20:00 Networking
How to apply
Spots are limited and every application is reviewed. Register here, then complete the short application:Open Google Form. Accepted participants will hear back by late November.
About Neuraffica
Neuraffica is a San Francisco company building AI from living neurons. We grow them on a chip, read how they compute, and use what we learn to make AI cheaper to run.