Junior ML for biology

Pay in the destination countries

Monthly, in US dollars. Each country is an Imora admission package — open one for that route's fees, language requirement and deadline.

What does the day look like?

A Bio-ML Engineer in private industry works 40 to 45 hours per week. There is 0% shift work and no routine on-call or night shifts. Model training runs overnight across cloud GPU clusters, sending automated alerts to Slack or email if an execution crashes.

Only 15% of the day is the glamorous part that outsiders imagine: training generative AI models that design novel therapeutic molecules. The largest share, 35%, is spent cleaning biological data, parsing coordinates from PDB and mmCIF files, normalizing assay readouts, and debugging tensor pipelines. Another 30% goes to training architectures in PyTorch or JAX, tuning hyperparameters, debugging CUDA out-of-memory errors, and running spatial validation splits. The remaining 20% involves MLOps, building Docker containers, optimizing GPU inference on Triton Inference Server, and meeting weekly with wet-lab scientists.

Where does this lead?