No, not using induction, just traditional resistive heating plus materials, shape and heater design to get even heating.

Preventing overshoot is a matter of the control method. We have a model of the system and use a predictive control algorithm to cut off heating just when we need to in order to hit the target without overshoot.

So is that to say if you introduce something to the cooking element, like water, the temperature will be wrong because the water unexpectedly cooled the surface?

We raise the temperature back up as quickly as possible without overshooting in thst case. During low temp cooking, you really only care about the maximum temperature, so you don't care much, it just takes a little longer to finish. But during sear of a protein, you just drive the temperature high enough that any loss from adding food or water to the plate is acceptable.

Do you individually calibrate the grills, or keep the same model parameters for all of them?

Our sensor elements are accurate out of the box, so in theory we shouldn't have to. But I'll admit, for our beta units, we still are calibrating each unit, largely for our readout electronics. We'll either fix this in the production model, or we'll have a highly automated calibrator.

So it's basically a temperature controlled soldering iron for food?

I try to eat as little solder as possible, even RoHS stuff :-) more seriously, getting a whole 10x10 plate to the same temp is trickier than just controlling a single point...