Translational ML Scientist

About the role

We’re looking for a Senior ML Translational Scientist to help turn Evvy’s microbiome data into diagnostic and prognostic tools that power our products. You’ll work with the world’s largest dataset on the vaginal microbiome paired with longitudinal symptom, treatment, and outcomes data.

Responsibilities

• Build the diagnostic and prognostic models that power Evvy's clinical and consumer products – including classifiers, risk prediction, biomarker discovery, and treatment response and personalization models
• Work hands-on with Evvy's proprietary dataset of shotgun metagenomics paired with patient-reported outcomes
• Own models end-to-end, from feature engineering and training through validation, calibration, performance monitoring, and the documentation engineering needs to take them from prototype to production
• Partner closely with leadership and Evvy’s R&D team to define what each model needs to do, what “good” looks like in performance and interpretability, and how to validate against real-world outcomes
• Translate model outputs into features and diagnostic tools that clinicians and patients can use
• Help shape Evvy's modeling roadmap, identifying where the dataset has unique leverage and what to build next

Requirements

5–8 years of post-PhD or post-MS experience building ML models on biological data; PhD in computational biology, bioinformatics, machine learning, statistics, or a related field strongly preferred (we're open to exceptional MS candidates)

Qualifications

• Genuinely strong in both machine learning and computational biology. You can hold your own in a methods discussion with ML engineers and in a biology discussion with researchers
• Industry experience shipping ML or computational biology models, not just publishing them. You've built things that ran in production and served real users
• Hands-on fluency in Python and modern ML frameworks (PyTorch, scikit-learn, and similar), and with microbiome-specific tools (QIIME2, MetaPhlAn, HUMAnN, or similar)
• Deep familiarity with the realities of microbiome data, including compositional data, sparsity, batch effects, and the difference between taxonomic and functional profiles
• A strong product instinct. You understand that the deliverable isn't the model, it's the feature or diagnostic the model enables
• Strong cross-functional collaboration skills. You can work effectively with clinical researchers, product managers, and engineers, and translate between their worlds
• A high agency, low ego. You move fast, decide with imperfect information, own outcomes fully, and care more about the work being right than about being right