As a guesstimate, in the past the best policy for reclaimed laptop cells has been to keep the continuous discharge rate below .5c max. They might handle more, but in general, those who obeyed this limit had success. In a laptop, the actual rate is way below .5c, so they don't put the high rate cells in them. They might be called 1c cells, but less than .5c is the rate you should aim for. And half that again, .25c, will work even better. So go big as you can, possibly by building one battery now, and another similar one later so you can run as low a discharge rate as possible later.
So your cells are supposed to be 2.5 amp hours ( 25000 mha). That was when they were new. Lets just assume they actually hold 2 ah on average now. That means .5c discharge rate would be one amp per cell.
20 amps controller, but in fact you wont pull 20 amps continuous, again, experience tells us that your setup will likely pull only about 10 amps max when up to full cruising speed.
So,, if you will pull about 10 amps continuous, you need to configure your cells in 10 p. ten cells or more in parallel. Then for 24v, I would suggest 7 cells in series, which would charge to 29.4v maximum. Likely 6 s would work too, but 7s 10 p is what I would shoot for. 70 cells. or if possible more, if you need longer range. This, assuming you get 2 ah per cell, would be a 24v 20 amp hour pack. That will be big enough for some decent range, more than 10 miles. Likely 15 miles, in fact. The above suggestion of 8p would be fine for new cells. But yours are not new, so you need to de rate your capacity in the calculation. Go with 10p or more. 10 p is still a relatively small pack anyway, for 24v.
Ideally, test every cell for internal resistance before assembly. Whatever the number is, and it might be poor, chuck out any cells that test a lot worse than the others.