I recommend Simon-Peyton Jones' classic "How to give a good research talk"
http://research.microsoft.com/en-us/um/people/simonpj/papers...And also some loose bits of advice from me:
* 10 minutes is not enough to present "all the material in your paper". 30 minutes is not enough either. So, don't even try. Even if you talk fast enough, the audience won't get it. I'd expect 10 minutes be enough for a high-level overview of the paper, but if you can include just one bit of technical information which you think is the most valuable bit, that would be brilliant.
* reading from slides is an atrocity, so don't.
* and having too much text on the slides isn't cool either (the same applies to bullet lists, unless you absolutely have to). If you do, expect many people not to listen to you while they are reading the slides. So, only include large bodies of text if it's some code. And even then, format it nicely (color helps, but don't rely on it to convey useful information).
* "pictures are worth a thousand words"... unless they are badly drawn. But a good picture is a great alternative to large bodies of text.
* practice your talk, especially if you haven't given many of them. That makes huge difference in how well you perform. And remember, a technical talk is still a performance.
* if you are worried that your mind would go blank once you take the stage, hard-script the beginning of your presentation and read from the paper. Eventually, you will warm up and won't need it anymore, but it's great to quickly overcome stage fright and the like (helped me a great deal).