The problem with an atmosphere hotter than lava is that very few materials are solid at those temperatures, and it's hard to imagine how a civilization could build an energy harvester without solid materials.
Everyday objects don't work like that. Saying ice cubes are cold is roughly equivalent to saying that they radiate less heat than the objects around them. If they radiated at "high intensity" then they wouldn't be cold anymore.
"A cold object that radiates heat at high intensity" is a contradiction.
[UPDATE] I don't actually know if current exoplanet observing technology is capable of detecting such a temperature gradient, but given what I know we can observe I'd be a little surprised if it couldn't. A planet-sized energy harvester would make a pretty big dent in the passive thermodynamics, and detecting that should not be too hard. And it would be Really Big News.
But when the rocks start to melt you have problems.
The good news is that you don't need to have the life arise under the same conditions that the technology exists. It's possible that the planet is inhabited by self-replicating tungsten-based technology that was created by life that arose somewhere completely different.
But the bad news is that we are much less likely to find the aliens than we are to find the descendants of the self-replicating robots they built millions of years ago. And the fact that we haven't found the robots makes it very likely that the aliens don't exist.
That being said, it’s never actually aliens in astronomy. So far, anyway.
I can't think of anything that would make a difference in the long run. Anything they do is still subject to the Second Law and the limits of Carnot efficiency. They can shunt a little bit of the energy off to the side and store it for a while, but it all has to end up as heat sooner or later.
Doesn't all the non-heat work eventually just become heat? Or am I misunderstanding your usage? Like a car with a solar panel still ends up radiating work as heat, by either air resistance (heat) or brake friction (heat).