Adrian Raudaschl

The way I think about product work still comes from medicine. On the wards, you figure out what's actually wrong before you act, you watch what's happening rather than what you assume is happening, and you don't lose sight of who you're there for.

Most AI products are optimised for speed and fluency. I'm more interested in whether a researcher can actually stand behind what the system produces: trace the evidence, see where it's uncertain, and decide what to trust. I think earning that trust is the hardest and most interesting part of the job.

I think the interesting thing about going from medicine to AI for research is that the core problem is the same: people doing difficult, important work need the right knowledge at the right time. The waiting room just got a lot bigger.