Early-stage biotech · AMR diagnostics

An early-stage biotech was betting on machine learning to read mass-spectrometry data and predict antibiotic resistance. It’s a field where the biologists know what the data means and the engineers know how to ship it, and neither can do the other’s half.

I built an initial version of the ML pipeline, and wrapped it in a CLI simple enough for a non-coder to run. I also built a demo that we took to hospital and HMO labs, where clinicians liked what they saw and gave us the feedback that shaped it.

The part I was proudest of was turning the domain workflows into AI agent skills that a non-developer could use. That let me coach the rest of the team into landing real commits through Claude Code: a biochemistry PhD who’d never touched Git, and a data scientist new to AI tooling.