Mad engineers with super engineering powers at that. I work in robotics -- in real life, the amount of effort necessary on the part of many engineers, some with Ph.D.s, to get even a simple robot like one of our AUVs assembled, integrated, and functioning with acceptable reliability is phenomenal. Yet Dr. Robotnik can build a whole robot army numbering in the thousands of many different types, a fleet of airships to deploy them from, and a chaos-emerald-powered world-shattering satellite -- all by his lonesome, all over Sonic's summer vacation, and all without having to worry about malfunction except via high-speed impact by blue hedgehog.
I believe you've uncovered one of the great flaws of standard testing procedures. The tests cannot test anything the developers don't think up. Robotnik has a big blank space where a blue hedgehog should be.
Despite his monumental engineering feats the guy didn't account for some pretty basic deficiencies. If I were him I would set up a test jig consisting of a cannon which fires spiked cannonballs at high speeds, and make sure my killing machine's armor could withstand at least that before deploying it. I would also take care that it never enters a failure mode that involves exposing its power core, CPU, or other delicate parts; and I would license Tony Stark's arc-reactor technology to power it, as the very rare mystical gems I would otherwise need tend to be either highly sought after by ambitious hedgehogs or guarded by fierce echidnas who would stop at nothing to ensure their safety and sanctity.
The Manhattan Project physicists occupy a dual space as well. Their ultimate hypothesis was "hmm, maybe building a bomb this way will work", and the experimental setup was "build it and try exploding it in the desert", which is a pretty engineeringy kind of science experiment. (Of course, there were plenty of earlier, more basic-physics experiments leading up to it.)
My favourite part of that experiment was that they believed there to be a non-zero chance that it would trigger an uncontrollable reaction that would ignite the atmosphere and end the whole world but did it anyway.
Nah, the LHC dudes think there's an actual zero risk. The LHC doesn't produce collisions of any magnitude greater than collisions that occur in the upper atmosphere (from what I understand). That was the ultimate debunk of things like the whole "microscopic black hole will eat the earth" thing.
The Manhattan Project physicists -- in particular J. Robert Oppenheimer -- joined Victor Frankenstein as archetypal mad scientists post-WWII.
In particular, fictional mad scientists tend to implicitly operate in the manner you describe: by the time they're building fully functional death rays with which to threaten the world, they've presumably carried out maybe tens or hundreds of experiments to see if the new form of radiation from a crystal meteorite energy source could be channelled and weaponized.
Noboby will tell a story about a mad scientist that's just a scientist and who can't build a giant death ray because no publication will accept his paper on it and he fails to get his department fund it.
Mad engineers are dangerous. Mad scientists are harmless.
Actually I suspect the origins of the term do in fact refer to scientists. Victor Frankenstein was in fact a scientist in the traditional sense (nor was he originally written as a villain).
But yeah, it's a lot easier to make a "mad engineer" villain than a "mad scientist" villain. Mad scientists tend to be more suited to being secondary characters-- somebody so obsessed with discovery that their goal of scientific discovery conflicts with the protagonist.
