We built robots that prepare fresh crêpes — now we need AI to be their eyes.
Our robotic crêpe station handles the full cooking process: pouring batter, spreading it across a hot plate, and folding the finished crêpe. But to do this reliably, the system needs real-time visual understanding of what's happening on the griddle. That's where you come in.
Your challenge: Build a computer vision system that can analyze images of crêpes on the cooking surface and determine: Spread quality — Was the batter evenly distributed? Is the crêpe circular and uniform, or are there thin spots, holes, or irregular edges? Position tracking — Where exactly is the crêpe on the hot plate? Provide accurate coordinates and orientation so the robotic arm knows where to act. Fold detection — Has the crêpe been folded? If so, how — in half, in quarters, into a triangle? Or is it still flat and unfolded?
Why it matters: Robotic food preparation demands tight feedback loops. A misread position means the spatula misses. A poorly judged fold means the crêpe gets mangled. Your model's accuracy directly translates to whether someone gets a beautiful crêpe or a sad pile of torn batter.
What to expect: You'll work with real images from our robotic setup, dealing with the messy realities of cooking — steam, browning variations, overlapping edges, and batter that doesn't always cooperate. Bring your best ideas in object detection, segmentation, or classification — and help us teach robots to make the perfect crêpe.