Machine Learning Scientist

Curve Biosciences™ ("Curve") uses Whole-Body Intelligence to monitor chronic diseases. We’ve created the first molecular blueprint of the human body by manually curating the world’s largest collection of comprehensively-characterized tissue samples into our Whole-Body Atlas. Trained on the clarity of our atlas, our Whole-Body Intelligence models identify chronic disease states through our Whole-Body Blood tests earlier and more accurately than other methods.

Our mission is to provide doctors the best intelligence for their patients and to alleviate the pain of chronic diseases by anchoring medicine in biological truth.

Our potential to transform chronic care is intertwined with applying our biological insights and machine learning methods to patient data. We’re seeking a Machine Learning Scientist who will collaborate closely with our computational scientists and Chief Scientific Officer to design, train, and deploy novel models for understanding disease biology from next-generation sequencing (NGS) data. To excel in this role, you possess strong technical expertise in the statistical and mathematical fundamentals of AI/ML tools, exceptional communication skills, and a cultural fit that embraces teamwork, adaptability, and innovative thinking.

You will operate at the nexus of machine learning research and real-world deployment, developing models that inform clinical and experimental decisions. This position is ideal for a proactive, detail-oriented professional who is eager to quickly integrate into a collaborative team, leverage existing knowledge, rapidly acquire new skills, and iteratively develop impactful models using in-house datasets.

• Design, train, and evaluate machine learning models (classical, deep learning, and hybrid) on large-scale NGS datasets.
• Develop novel modeling approaches for extracting disease-relevant signals from high-dimensional biological data.
• Collaborate closely with computational and experimental scientists to ensure models reflect biological reality and inform assay design.
• Translate model outputs into actionable insights that guide experimental and clinical decision‑making.
• Contribute to publications and present work at leading AI/ML venues (NeurIPS, ICML, ICLR, AAAI, and similar).
• Write clean, reproducible, and well‑tested code following best practices in scientific computing.
• Promote a culture of scientific integrity, growth, transparency, collaboration, mutual respect, and fun, while contributing to our goal of improving patient lives.

• A collaborative, inclusive team player who thrives in interdisciplinary environments.
• A clear and thoughtful communicator, comfortable explaining complex ideas across disciplines.
• Curious, impact‑driven, and motivated by scientific discovery.
• Humble, open‑minded, and eager to learn.
• Passionate about applying data and science to improve patients’ lives.
• PhD in Computer Science, Statistics, Computational Biology, or a related quantitative field (or Master’s with equivalent research experience).
• Track record of first‑author or significant contributions to publications applying machine learning to biological or biomedical data.
• Strong hands‑on experience with PyTorch and the scientific Python ecosystem (Num Py, Sci Py, Pandas, etc.).
• Demonstrated experience designing, training, and rigorously evaluating deep learning models (ablation studies, failure analysis, and interpretability studies).
• Strong intuition for modeling trade‑offs, including when to apply simpler vs. more complex methods.
• Experience with or knowledge of ML infrastructure engineering best practices, including GPU‑based training workflows.
• Experience with cloud environments (GCP, AWS), high‑performance computing, and version control (Git/Github).
• Able to effectively perform scientific literature reviews that drive insights.
• Experience contributing to and maintaining deep‑learning codebases, with a high bar for engineering quality, reproducibility, and testing.
• Ability to work from our San Mateo office at least 2‑3 days per week.

• Experience in building DL models for genomic data, with knowledge of state‑of‑the‑art genomic…