Struggling to present results from long papers in short time slots

Attempting to fit the material of your paper into a talk is a common mistake: even for shorter papers, there is typically simply too much to include all of the significant details.

Instead, I recommend thinking of your talk as an advertisement for your paper. Your goal is to present enough of the key interesting material to be able to convince somebody that it is worth actually reading your paper.

Once you accept this as your goal, the length of the talk does not actually matter. As an exercise, you might even try making just a single slide for giving a two minute talk (as one might in a "lightning talks" session).

What, then, should you put into whatever time you have available? Think of the story of your work, as opposed to the content. Your abstract may help you here. As a starting point, I would suggest this as a general framework suitable for most talks:

• Set up the problem, explaining why somebody should care about what you are doing.
• Sketch the approach that you are taking, explaining why it is reasonable.
• Point out some critical insights that were necessary to make the approach work.
• Show your key results and explain why they are significant.
• Come all the way back to the set up and explain the progress that has been made on the problem.

Depending on the details of your work, there may be some differences, but this may at least help you get started sketching things out.

My concern with this approach is that I may not be taken seriously; there seems to be an unspoken norm in my field that young academics ought to present painful details in order to prove how smart they are.

There are all kinds of norms that are wrong, and this is one of them. Painful details are just painful and nobody can follow them during the talk, anyway.

Without knowing your work, it's hard to give specific guidance but the goal of your talk should be to convince your audience that your work is interesting. Proofs are usually boring. Your framework might be interesting. Your "cool results" are certainly interesting to you so you need to make sure that your audience even understands why the area they come from is interesting. But everything needs to be at the level of explaining the big picture and roughly how the system works. Don't be afraid of saying things that aren't quite true, if the precise statement is just too long and hard to follow. For a trivial example, it's OK to pretend that all prime numbers are odd, as long as you mention that this isn't quite true, and as long as 2 being prime doesn't mess up your whole argument. Keep things mostly about intuition rather than about fiddly technical details.