Unreasonable battery charger

[Zyeborm]

Morning all, this does belong here honest 😂.

I'm thinking of making a 12s lithium battery charger as a bit of a stretch project for myself. Digital design I can do reasonably well but I'd like to improve my analogue chops.
As such I'd like to make it unreasonably accurate in terms of measuring cell voltage and capacity and do so using as few off the shelf modules as feasible without going to the point it doesn't actually get built.
I'm focusing on the cell voltage measurement stage at present.
What I'm thinking is this would be a nice place to try out making a dual slope ADC.
Told you it was unreasonable right?
I'm thinking of using a large value flying capacitor (PP or C0G)  in a ladder to capture the cell voltage and bring it down to a ground reference, buffer that then feed it into the ADC.

I'm hoping that many of the difficulties with dual slope could possibly be avoided in this application.
Because our battery voltage is always at least say 2.7 volts we could avoid operation around the zero volt transition that from what I read seems tricky. So instead of a negative reference use a 2 volt reference.
Speed is obviously not a huge deal, if I can get 12 stable readings a second it'd be fine. So we can keep the speed of everything down nice and low.
10 measurements per cell at 12 cells nets a neat 120 hz to match double the power line frequency here. So I could trigger one reading on going high and one going low to balance that out. 120hz seems low enough that frequency response of amplifiers shouldn't matter too much?
It will presumably need a much chunkier cap for the dual slope than is generally used for this kind of thing.

I have no fixed design on anything at this stage, and I have no idea about specific parts that would be good for this build.

I'm after general advice from those more experienced especially with dual slope and analogue in general.
Are there any parts you would suggest for opamps and comparitors? Is it slow enough that I can reuse the same opamps as comparitors?
Is it so slow that all those pesky second order analogue effects will render the output worse than the adc of an esp32 with some 5% resistors in front of it 😂

Is a MOSFET based ladder and flying cap going to be too leaky to get useful measurements?

Thanks so much for all your time, this is my first foray into participating in volt nutting rather than just living vicariously through Marco, hopefully it's a lasting one