Hi @slensgra, this progress is really cool to see! 
I enjoyed the writeup, and particularly like the wifi access and display. It’s worth noting that while an aluminium tube would definitely help with keeping things cool, it would also make wifi and the diver-friendly display harder to get working. The display may be able to be replaced with a few penetrated indicators, or perhaps just one with an RGB led or similar, although lights at the back are likely not quite as convenient as being able to just look through the sides. Perhaps it would also be possible to have an external wifi antenna via a penetrator, although I’m less certain of that (as a side-effect, that could be placed such that it’s out of the water without the vehicle needing to be lifted when it’s at the surface).
Have you considered using some sort of leak sensor for better handling of leaks? If you’re expecting that leaks may occur and you don’t want to stop operation for a slow one, perhaps you could allocate a small section at the bottom of the enclosure with a highly absorbent material in it (to avoid splashes/sloshing), and potentially have leak detection above and in that section, to detect when a leak starts, and when it gets severe enough that switching off is essential.
More generally, given the ideas behind this system, we’d love if you could share some thoughts on autonomy in marine robotics: