Scientist, Machine Learning (Human Physiology)

Flagship Pioneering•Cambridge, MA

6d•$136,000 - $211,750

About The Position

FL127, a privately held, early-stage company in the Flagship Pioneering ecosystem, is developing a novel device platform through a hypothesis-driven experimental program. The company is building a lean team of scientists and engineers working across hardware, software, and physiology, with AI agents integrated into daily work. The Scientist will connect physiological signals measured by the platform to what those signals indicate about a person. The role grounds measurements in physiology, helps design human studies to test their meaning, and builds models relating platform measurements to reference measures and outcomes in small human cohorts. The imaging and signal-extraction team produces the measurements; this role begins with those outputs and collaborates with that team to design studies that assess both measurement quality and interpretation. Team members use AI agents for literature synthesis, study design, data analysis, and code. The salary range is $136,000–$211,750. Compensation depends on qualifications, skills, competencies, and experience. FL127 offers healthcare coverage, an annual incentive program, retirement benefits, and other benefits.

Requirements

PhD in biomedical engineering, physiology, bioengineering, biophotonics, or a related field, or an MD/PhD, with substantial applied machine learning experience.
Deep knowledge of human physiology in an area such as microcirculation, hemodynamics, tissue optics, skin biology, or cardiorespiratory physiology.
Experience taking a measurement or digital biomarker through human studies, including endpoint selection, reference measurements, IRB-governed protocols, and agreement analysis.
Experience with statistical modeling on small human datasets, including within-subject and longitudinal designs.
Proficiency in Python and the modern machine learning stack, and comfort working in an existing research codebase.
Imaging and signal-processing background sufficient to reason about how calibration, registration, and sampling affect a derived measurement.
Ability to work independently in an early-stage, ambiguous environment and make decisions with incomplete data.

Nice To Haves

Published or applied work in remote physiological sensing, digital biomarkers, or wearable or contactless measurement.
Experience turning research results into a product feature, ideally at a consumer device company.
Familiarity with skin optics models, such as Monte Carlo simulation, diffusion theory, or chromophore decomposition.
Experience modeling individual baselines and change over time.
Experience in a venture-backed startup or early-stage research environment.

Responsibilities

Ground measured signals in physiology, including what they reflect, what confounds them, and which physiological states they may indicate.
Work with the product and market team to prioritize applications for study by considering scientific plausibility, study cost, and user value.
Co-design human studies with the imaging team, including endpoints, reference measurements, and protocols.
Build and evaluate models relating measurements to reference measures and outcomes in small cohorts, accounting for confounders, leakage, demographic stratification, and sample size.
Set the evidence threshold for findings before they influence the product, and communicate what the data supports and where it falls short.
Review extracted signals for physiological plausibility and suggest improvements to the measurement pipeline.