Prima Mente's goal is to deeply understand the brain, to protect the brain from neurological disease and enhance the brain in health by generating data, building brain foundation models, and translating discovery to real clinical and research impact.
Requirements
- Strong background in machine learning or a closely-related field (e.g. deep learning, statistics, optimisation).
- Demonstrated experience training and evaluating modern ML models (e.g. transformers, diffusion, graph models, sequence models).
- Solid software skills in Python and at least one major ML framework (PyTorch, JAX, or TensorFlow).
- Experience designing and running non-trivial experiments: controlling for confounders, building robust baselines, and doing thorough error analysis.
- Ability to write clearly – whether in code comments, research docs, or papers.
- Experience with large-scale data (e.g. 100B+ tokens or equivalent) or distributed training.
- Background in computational biology, genomics, epigenomics, neuroscience, or related areas.
Responsibilities
- Design and implement ML models for large-scale biological data, from pre-training to task-specific fine tuning.
- Partner with biologists, clinicians, and data scientists to translate biological and clinical questions into tractable ML problems.
- Run end-to-end experiments: dataset curation, training, evaluation, error analysis, and iteration.
- Develop and refine evaluation suites for robustness, generalisation, and clinical relevance (e.g. across cohorts, sites, populations).
- Explore multi-modal and multi-task training across genomic, epigenomic, transcriptomic, proteomic and clinical signals.
- Perform in-depth model analysis to extract mechanistic or biomarker-level insights, not just metrics.
- Collaborate on papers, internal memos, and external communication of key research results.
Other
- Motivated by advancing human health through AI, especially in neuroscience and complex disease.
- Deeply curious, with a habit of reading papers, prototyping ideas, and stress-testing your own assumptions.
- Comfortable doing real engineering work in service of research – but see yourself first and foremost as a researcher.
- Enjoy collaborating across disciplines and explaining your work to people with very different backgrounds.
- Able to stay with hard problems for a long time, and to make progress even when the path isn’t obvious.
