How do I get better at giving talks/presenting research?

One important thing, for both casual talks and more formal talks, is your notes. Too many details and you can't use them without interrupting the flow of the talk, too little and you can get lost.

After some time, I converged to the following solution. I usually prepare one or two pages of main notes, not too densely written, with the main items I want to speak about and large signs for important things to mention at some point. Beyond two pages, I just end up caught in the moment and diverge from the planned path and improvise too much.

Usually there are some points where I could need a complicated formula, a precise statement, or a question might occur that needs one of these in order to be answered; then I simply make a mark (circled capital letters of color) where they might be needed, and write down the details along with the same mark on a second set of more detailed notes. Then I can refer to these when needed.

Another, general advice: be careful of what you do not like in other's talks, think about why, and be sure not to do that. The flaws of talks are often broadly shared, and this can only happen because we do not realize we do exactly what we blame others for.

The generic answer to this is probably "practice", but in the interest of giving you something more concrete to work with, I'll share what works for me personally:

When I make slides, I will usually formulate out what I say along with a given slide in my head or even under my breath. It helps me figure out when the slide isn't quite jogging my memory correctly, or when the organization doesn't follow my natural train of thought. I'll also usually notice if there's something I keep forgetting, and I can do something about it - practice that slide a bit more, add a reminder on it.

That way, I can easily get in 6-7 "rehearsals" of my slides just during the drafting process. Of course it's not as good as a real rehearsal, but if you don't have the time it's at least something. What it doesn't do as well is transitions between the slides, and I'll occasionally forget what order I have them in. I re-read my slides just before the presentation and also try to have my laptop screen show me a miniature of the next slide if possible to help with that.

Honestly there is no better means than to practice: but setting and audience matters as well, here are some pointers:

• Location Counts: Practice at the venue if you can, or a similar location. Practicing at home is different than on-stage.
• Know your audience: Practice with subject experts and laypeople. The information you give will need to be customized to the audience. If they are experts then you can skip 'basics', if they are not then you need to be aware that what's easy for you may not be easy for them.
• Content: Slides are great for visuals, not really for text. When you present, you are telling a story. Having the audience read the story does not make for a compelling storytelling presentation.
• There is no script: Trying to adhere to a script means that you will often hesitate and wonder if you are on-script or off-script. Instead, concentrate on talking points that you need to cover. Your expertise and experience on the research will fill in the actual wording for each bullet point.
• Dress the Part: Your clothes will give you confidence. Although it may be psychological, the idea here is that formal wear will allow you to psyche yourself into thinking that you will do well.
• Umms and Like: These verbal fillers detract from the message and are a sign of nervousness. The audience can't tell if you are nervous or not if you take your time and speak slowly.