At no point during its 4 year journey does the Kassandra "flip" over. Any reaction engine powered space ship would accelerate for around half of its journey (or until terminal velocity is reached), then turn around and decelerate to arrive at its destination at a sensible velocity.
Then idea of having vast empty spaces on a star ship is ridiculous, and having them pressurized and with snow flurry and a star ship accelerating for years with what looks much like rocket engines even more so. Especially on a ship with the drive going continuously to ac/decelerate at what appears to be around one g, not heating but getting rid of the excess heat would be a technical challenge. And even with an imaginary, thought-experimental drive that could transfer all the energy from hydrogen-fusion to the produced He and turn all of that energy into momentum (which is far more than the laws of physics will ever allow), to accelerate some mass for four years at 1 g, the ship would have to start with millions of billion times that mass of fuel.