Most engineers are what we call "instrumentalists," which means that they don't care if theories are true or not; only whether or not they can be applied to solve problems.
But if you're still curious, here are few simple arguments for both sides (realist and anti-realist):
Antirealist:
--> scientific theories aren't true and their entities don't exist
Pessimistic Induction: Since all our past theories have all been proved wrong, our new theories must also be wrong (because they will be proved wrong in the future).
Science is non-cumulative: Every time there is a new theory, it completely replaces the old theory because the old theory is now seen to be wrong.
Realist Rebuttal: The above argument considers theories to be indivisible wholes. Real theories can be broken into chunks, and we can keep the correct chunks when new evidence presents itself. In this way, science is cumulative because a chunk is always be passed from one theory to the next.
Underdetermination: Given enough creativity, there are infinitely many theories that fit any set of evidence, but differ when used to extrapolate. Therefore the theory we are using is likely false.
Realist Rebuttal: The philosophers that said this were never able to give any examples.
Realist:
--> scientific theories are true and their entities do exist
Miracle Argument: The only way you can explain the success of science, is by realizing that the theories and what they say is true.
Anti-Realist Rebuttal: It is a competition of the fittest. There are initially many theories, but only the strongest survive. They however are still not completely true.
Overall:
It is a difficult question because we can never prove any scientific theory to be right; you can only prove it to be wrong. I do however think that electrons exist, but I do not think we will never know everything about them. It was the height of human ignorance when Lord Kelvin said that there are only two things left to be discovered in science: black body radiation and the Michelson/Morley experiment. And since we will never know everything, doesn't this mean that we will always be working with approximations (as opposed to truths)? Remember, Newtonian mechanics (F=ma) are technically wrong, but they act as good approximations for 99% of the work out there. Only in certain cases do we need to consider relativitic physics or quantum physics.