The problem with planning to use a guy for "low-leverage" innings is that those innings will come at inconvenient times.
Say your starter has a bad game and you're behind 6-0 before he can even get out of the second inning. It'll be low-leverage from here through the ninth, unless your hitters have some kind of outburst of their own. You bring in your mop-up/long reliever, and he goes 3 1/3, say. No big rally, so you're still down 7-2 after five, say, with three or four innings to go (away vs home). Those innings will be low-leverage too. But the guy you tagged for those is done. Now you bring in a sequence of your one-inning guys, presumably in reverse order of ability and seeing if you can squeeze an additional inning out of one, but by this point it's not going the way your low-leverage plan envisioned.
And then the next day, you're winning 8-3 in the eighth, and your long-man low-leverage guy isn't available to mop up because he's resting.
And then you have an open date on the schedule, and then you encounter a string of five close games in which you wind up bringing in your low-leverage guy to face the #7 batter to start the inning and he walks a guy or two and suddenly he's facing the top of the order and you grit your teeth and you hold your breath and then possibly wind up taking him out anyway and burning another arm you didn't want to.
There exist low-leverage situations - plenty of them during a season in fact - but as Chief says, there are no low-leverage relievers anymore. More precisely, no low-leverage roster spots in a major-league bullpen. The usage of starters has changed since I was a youth, and this is a consequence.
I still wanna keep Castellano, but he has to perform at least as well as the waiver-wire dreck that is available.