I recently landed in bioinformatics coming from a semi-related semantic web project (coincidently at Stanford, in BMIR). It's one of the few academic research environments I've seen that actually use cutting-edge production systems (prototypes aren't enough). It's a ton of fun and has direct applicability to startups, especially areas that are working on huge datasets and enormous graphs.

I work on BioPortal, here's a sampling of the technology we're using:

Ruby

Rails & jQuery (UI, REST client)

Java (Spring, Hibernate, Protege, Lexgrid REST service)

mySQL

VMWare

Python for some metrics calculations

For our next iteration, we're going to be looking at using document stores for our ontologies, deploying a triple-store, and possibly some MapReduce for processing some of our ontologies (which can currently take 2-3 weeks to get through our entire workflow).

I'm happy to answer questions if people have them, or if you're interested in getting into the field and want some advice just let me know.

Why aren't you using appengine?

What exactly is the point of triple stores? they're always much slower than a proper schema.

Do you use genetic programming in bioinformatics? if not, why not?

Is it possible to make profits in bioinformatics (ie. not be government funded) ? Specifically, are there free datasets that you can "infer" new information from using machine learning/genetic programming and sell this information for big bucks?

Interestingly, you could pretend the 'Bioinformatics' was not in the title and it would still be an interesting article. Though perhaps somewhat redundant for most folks here.

good point, it's really an introduction to 'how to get disciplined and organized in your code hacking'