People are fallible on the best days, assuming everyone did their very best from nuclear physicists to construction workers, mistakes are made. You take steps to reduce the risk. Research gets review. Engineering schematics get review. Construction gets inspection. Still some mistakes will get through.
And people always act their very best all the time right?
You can even have a perfect design, perfect construction, that is mismanaged years after it's built, after the original engineers and bureaucrats lose control.
The same people problems apply to basically every human endeavor, but nuclear's capability to cause accidents that have a lasting impact is pretty scary. You don't feel even a twinge of existential dread when you think about? If you don't, then I don't think I want you working on a reactor.