Any possibility of using closed-loop feedback from an IR sensor to control temperature through your board? I'm working on a system that does this with a flame right now, but an all-electric solution might be neat too.
Andrew and I have worked on that a bit. The most obvious solution, an optical pyrometer, won't work because brass has such a low thermal emissivity constant (0.04 for unoxidized brass) and it varies so widely with surface finish (in the 0.40 to 0.60, depending on surface finish and type of contamination. If you always used brass fresh out of the tumbler then perhaps optical could be made to work.
Another thing that we've looked is a thermocouple spring-loaded against the case mouth. There are two problems, one easily managed and one not so much. The easily managed one is that RF from the heater tends to get back into the T/C controller. I pretty much stopped that by wrapping the lead wire a few turns around a ferrite toroid that I had handy.
The other problem that at least one of the wires in the most common types of thermocouples is magnetic. That means that it absorbs induction heat and does so better than the non-magnetic brass. A partial solution is to use T/C wire fine enough that it doesn't appreciably absorb energy. The downside is that this extremely fine and fragile wire is not durable.
We've both pretty much decided that a purely time-based system is good enough. Andrew has spent a lot of time with the Templaq verifying this. Unlike with flame heating, the induction heat is on and off instantly and so can be controlled very precisely. The only compensation that would be needed would be for changes in line voltage. A variac is a simple solution to that problem. Andrew is using an Arudino as the controller so line compensation can be easily added, if he isn't already doing it.