The presentations at a SciPy Conference are very cross-disciplinary. Some Presenters talk about interesting things they have accomplished in their scientific domain using Python. Some people present the latest developments in a particular open source application or library they are releasing. Still others talk about particular problems encountered with performance, interaction, etc. and how they solved them. Some speakers use the NumPy and/or SciPy libraries, while others are just using Python for a scientific task or workflow.
If you're interested in seeing what's happening with scientific computing with Python, this is the place to come. Many attendees are implementing Python in research (~55%), education (~20%), commercial applications (~25%), or just showing up to see what all the fuss is about.
A strong programming background is not generally necessary, but presentations will certainly use lingo and discuss concepts that might be unfamiliar. Speakers are generally kind enough to give good domain-specific background so that a technically proficient layman can follow what they've done. Tutorials are fantastic and target a range of background experience (see the Tutorials list for more information on these).
The downtime between talks is some of the most productive and interesting interaction at the Conference. It allows people to connect with old and new friends and peers and to dive down into details of what speakers and other attendees are up to with their scientific Python development.