Senior Machine Learning Scientist – Bioinformatics, Python, ML – USA, Remote

Key Skills/Responsibilities:

The primary priority is for the candidate to have publication‑backed experience, specifically in:

• Predictive ML on biological data (classification modeling)
• Genotype/feature → phenotype modeling
• Method development or algorithmic contributions
• Model interpretability and generation of biological insight

Technical Expertise

• Strong background in machine learning or statistical learning with substantial hands‑on experience developing classification models
• Experience working with high‑dimensional, sparse biological or omics datasets
• Strong proficiency in Python for end‑to‑end machine‑learning workflows
• Demonstrated experience designing validation strategies and assessing performance under significant class imbalance and limited sample sizes

Scientific Rigor

• Clear understanding of model limitations, uncertainty, and overfitting risks in real‑world biological datasets
• Experience delivering machine‑learning analyses intended to inform research and internal decision‑making
• Experience making principled methodological recommendations in the face of incomplete or noisy data

Preferred Qualifications

• Experience working with biological sequence data or genotype--phenotype analyses
• Experience with interpretability or explainability approaches applied to biological machine‑learning models
• Background in pharmaceutical, biotech, or regulated research environments

Machine Learning Model Development

• Develop classification models to analyze curated genotype--phenotype datasets
• Apply appropriate modeling strategies to predict viral sensitivity or resistance based on sequence‑derived features
• Implement training, validation, and hyperparameter‑tuning workflows using predefined datasets
• Evaluate alternative feature representations provided by the bioinformatics team and assess their suitability

Model Evaluation and Robustness

• Assess model performance using metrics appropriate for imbalanced biological datasets
• Evaluate robustness across data splits, phenotype definitions, and successive data releases
• Identify failure modes, instability, and limitations, and document their implications
• Document modeling assumptions, trade‑offs, uncertainty, and limitations in a reproducible and transparent manner

Interpretability and Insight Generation

• Provide interpretable summaries of model behavior, including feature importance and consistency of signals
• Identify amino‑acid positions or features that recur across models or resampling strategies, while highlighting where signals are not reproducible
• Clearly document and communicate findings, assumptions, and caveats within the bioinformatics team