Well, start by remembering that the software is going to be a tool not the product. Therefore it has to work correctly in terms of provable results but the longer term maintainability is less important. Slick, aside from publishable data, is less important as the users are domain experts.

Try to understand the research work being done, “going to school” on the material as needed. Scientific chauvinism can be strong at times and people who don’t know the jargon can be (or will be) dismissed as technicians. The fact that you’ll know far more about software development won’t matter as that’s not of real interest to them. Fair? No, but so it goes..

Hint: it’s pyridine not “benzene with an N in it”.

Certain fields have been well served by Foss software and even tech books. Make sure you know what’s already there in case it can be adapted or used as test tools. Understand why what you’re doing must be different so you can defend yourself. There might be commercially available options but academic research isn’t usually interested in spending that kind of money. Pharma research is totally different about this, although cheap bastards are there too.

You’re out to prove you’re more valuable than a second year grad student. Be helpful, productive and be willing (and hopefully able) to push the project forward.