If you haven't already seen step two or one, rectify that before continuing…
So for the electronics you have again many options. First, you have to decide which sensors and where you place them. If you made it this far, you don't need me to tell you what to do… but I would mention that you may have the sensors measure the hammers, the keys or both (the latter is my choice). There are many kind of choices you have for sensors. Probably you will not want contact one, so either optics or hall effect will be your choice. The latter are discussed in https://pianoclack.com/forum/d/238-another-diy-controller-project whereas the former are discussed at https://pianoclack.com/forum/d/243-diy-piano-sensors -- More details on the sensor comparison here
If you go "optics", like I am doing, you have two main avenues: comparators, like the original Cybrid; or ADCs like piano-conversion. The ADCs will give you more flexibility, the comparators might be easier to design and build fast enough. However even a cheap Raspberry Pi Pico has an ADC which can run 500ksample/s (and could be multiplexed among up to 3 channels, giving 166ksample/s per channel). Not that I am advocating using that board: too many are needed, a nightmare to track them and having them talk with each other…. Other boards have much better ADCs, are only marginally more expensive (per channel), and provide either many more bits (16 bits) at a marginally faster 700ksample/s (to be dived by many more channels, though) or they can be configured with the same bit depth as the Pico (12 bits) and provide much faster total data rate. In any case the lower bits of the ADCs are garbage anyway, so the 12 bits of 16-bit ADC in the ST board might be more usable than the 12-bit of the Pico (which has only about 9 usable bits, as described in its datasheet).
The easiest thing to do with the ADC data is to have comparators in firmware (on each ADC boards, not on the controller as I did merely for testing purposes) and trigger MIDI note-on and note-off at appropriate levels. As needed, the firmware may do other things with the data, which would be very difficult with the hardware comparators. A possible "middle ground" alternative would be to use a DAC as the reference for the comparators, rather than a fixed voltage set by potentiometers as the original Cybrid did.
Stay tuned for more!