I shall try with a few comments 
Be aware, that this high DC voltage can be a challenge. If a fault cause a short circuit anything, the battery may be able to support very high currents, and it causes arching and fire risks. So you need good sand fuses in series with the battery to prevent some nasty events.
I would avoid the extra common shunt resistor, because it introduces increased self inductance of the ground current path. You are able to measure the currents in the three other shunts and add the results to provide the same signal. I do not like the self inductance in ground from the big electrolytic capacitor to the far away half bridge. In previous design, you also had a ground path on the other side of the PCB, and it helps. It would be better also to have two electrolytic capacitors and one on each end of the PCB to reduce this self inductance. It may cause a longer physical PCB, but the height of the board may become less overall.
I think you will like to be able to measure +/- 15 Amps with the shunts. This is 0,15 V on the 10 mOhm shunt. You cannot choose less amplification of the INA181 than 20, so it causes +/- 3 V. But you only got a range of +/- 1.65 V with the 3.3 V power supply. Therefore I suggest shunts of 5 mOhm.
I think you like to have a fair common mode rejection, CMRR with this INA181. It is guaranteed to be above 84 dB or 1:14.000 with zero Thevenin impedance on the reference input. But R19 and R62 makes a Thevenin resistance of 15 kOhm. The internal resistor is 500 kOhm according to page 22 in data sheet. Therefore the CMRR is reduced to about 33 or about 30 dB. I suggest you reduce the resistors to perhaps 1 kOhm. Another possibility is to make a common center voltage reference supply with close to zero impedance.
I am not that experienced in PCB design to ESCs, so I cannot say that a two layer PCB is sufficient or not. I know, that the signal integrity is very much improved by a main ground plane in the PCB. With four layers you will also be able to reduce the problems related to the self inductance of the shunts, that cause high voltage transients on the measured voltage on the shunts. I explained a bit about that in my other recent post tagged current_sensing.
https://community.simplefoc.com/t/shunt-current-measurement-considerations/7103
It also explain the need for some RC filtering here. If you got layer 2 as ground and layer 3 as power, two electrolytic capacitors are not an issue anymore.
I tried to look into choice of power mosfet. The IRF540 is an old design, but it is also widely available. But I like to have a bit lower power loss to avoid a heatsink. So I suggest a mosfet with about half RdsOn. It could be IPP180N, with about same price or IRF8010.