Senior Manager, Data Science

BMS Digital Health is seeking a Senior Manager, Data Science to build and deliver hands-on, code-first analytics and algorithm development using wearable and sensor-derived longitudinal data. This role is for a data scientist who thrives in the details-owning work end-to-end from raw signals to validated outputs-spanning time-series QC, preprocessing, artifact handling, imputation, feature engineering, and modeling across accelerometry/actigraphy and cardio-respiratory signals (e.g., HRV, SpO2). The ideal candidate enjoys writing production-quality Python in orchestration environments, applying rigorous validation, and collaborating across internal and external partners.

This is a highly hands-on individual contributor role. You will spend a significant portion of your time coding, debugging, reviewing PRs, and building reproducible pipelines and models.

What You'll Do (Hands-on Responsibilities):

• Build and maintain Python pipelines for wearable time-series data, including:
• QC, preprocessing, and sensor artifact removal
• Imputation (baseline through advanced methods) and feature engineering based on clinical concepts of interest
• EDA and signal characterization for accelerometry/actigraphy, HRV, and SpO2
• Signal processing and signal detection
• Develop and validate models for longitudinal sensor data using:
• Frequency / time-frequency representations, digital filtering, and representation learning
• Quantitative characterization of physiological and clinically meaningful measures provably associated with disease progression or subtyping.
• Deep learning approaches (Transformers and/or ensembles) with model explainability techniques where appropriate
• Apply statistically rigorous approaches to repeated-measures data:
• Longitudinal statistical modeling (e.g., mixed effects / hierarchical models)
• Study-appropriate strategies for within-subject dynamics and missingness
• Implement strong evaluation practices and reproducible research standards:
• Nested CV, LOO, and/or OOB methods where appropriate
• Reproducible experimentation, documentation, and well-structured codebases
• Collaborate actively with internal stakeholders (clinical, stats, engineering, product) and external partners / third-party analytics providers, including QC and validation of vendor-derived outputs.
• Contribute to team excellence via code reviews, technical mentorship (scope depends on level), and raising engineering rigor.

Required Qualifications:

• PhD (preferred) or MS with strong experience in Data Science, Biostatistics, Biomedical Engineering, Computer Science, or related field.
• PhD 3-5 years, MS 6-9 years, prior experience working on digital health initiatives within pharma industry, medical devices etc.
• Demonstrated hands-on experience with time-series sensor data, including:
• QC, preprocessing, artifact handling, imputation, feature engineering for accelerometry/actigraphy
• Experience with HRV and/or SpO2
• Strong Python skills with evidence of shipping code:
• Clean, testable code; object-oriented design; modular pipelines
• Git/version control, code reviews, and collaborative development practices
• Experience with longitudinal statistical modeling for repeated measures data.
• Proven ability to translate analytical work into clear deliverables and communicate results to technical and non-technical stakeholders.

Preferred Qualifications (one or more of the following):

• "Navigational physics" experience for movement data and/or biomechanical analysis (a plus): quaternions, Euler angles, dead-reckoning, orientation/heading estimation, etc.
• Familiarity with sleep analytics and/or circadian cosinor modeling (or willingness to learn open-source libraries).
• Experience managing or integrating third-party analytics (e.g., actigraphy QC workflows) and validating vendor outputs.
• Experience with scalable compute and deployment patterns, including:
• AWS experience; multi-GPU instances and parallelization for model training/inference

Why Join Us:

• Work at the intersection of digital biomarkers, machine learning, and clinical development
• Partner with multidisciplinary teams to advance analytics from real-world and trial-based wearable signals
• Build practical, validated solutions with high impact-and write the code that delivers them

We hire for skills and capabilities, not just credentials - if this role excites you, but doesn't perfectly match your resume, we encourage you to apply anyway.