tbf, the only thing he set a strict deadline for was the prototype. Something that is never going to actually have any passengers, and no cargo beyond what's used to stress-test the weight limits. And sounded like he only want to know it works at all, so prototype also don't really need any turns. Meaning it can't really derail unless there is an inherent flaw in the design.
Honestly, I suspect even I could build a 1km long railroad with prototype-train in 3 months, alone. All I would need to do is source some standardized wood-pieces, cut it to length (2 short and 2 long - likely the long is just the original pieces, and the short is a quarter of the original piece), nail the short and long ones together such that it forms a straight rectangle of specified width, then carry and place after the last one (likely also nail into the dirt). Then get 2 rods, 4 circles, drill holes and glue together. Then put a simply frame ontop of the 2 axles. Then its power would come either from going downhill, or from having a leather/rubber band between one axle and a crank, or simply from attaching a pre-made engine (maybe a bicycle type of engine, aka foot crank?).
1km is long, yes. But is it 3 months kind of long? Doubt it. Particularly as measuring-efficiency, circular-sawing efficiency and nailing efficiency would pick up after the first month.
Then once that is done, the development of the real product can start. Everything from a magic-power based engine (likely the biggest issue), to proper standardization, ideal shape of wheels and rails, and other parts of the mechanisms.
edit: (t be clear, I am not talking about having a tunnel too. I doubt the prototype has a tunnel, since they already know how to dig tunnels)
