Senior Data Scientist, Medical Imaging
About this role
Heartflow builds an AI-driven, non-invasive cardiac test supported by clinical guidelines that creates a color-coded, 3D model of patient coronary arteries to assess blood flow and diagnose heart disease. The team is expanding its pipeline of medical technology solutions to improve precision heartcare worldwide.
This senior team member will drive data-centric artificial intelligence initiatives by examining how diverse real-world imaging conditions impact deep learning algorithm accuracy. Responsibilities include building analytical pipelines to quantify dataset distributions, exploring methods to harmonize complex image variations, and translating insights into data curation requirements for algorithm development.
Candidates who thrive in this environment possess deep expertise in medical imaging data structures, strong statistical analysis skills in Python, and hands-on deep learning experience for computer vision tasks.
What you will do
- Build robust and reproducible analytical pipelines over population-scale data to characterize dataset distributions and model vulnerabilities across diverse patient populations.
- Explore, develop, and validate methods for harmonizing complex data-related factors and image variations.
- Translate findings about data variance and model behavior into actionable data curation requirements, training-time robustness strategies, and architectural recommendations for downstream algorithm development.
- Develop annotation protocols for algorithm training and validation for internal product development and FDA submissions.
- Partner closely with research scientists, machine learning engineers, systems engineers, process engineers, product teams, and regulatory teams.
- Derive and present clear analyses from messy data to drive decision-making and provide consumable artifacts for other functions.
What you need
- Hold a Master's or PhD degree in Data Science, Computer Science, Medical Image Analysis, Statistics, Biomedical Engineering, or a related quantitative field.
- Bring at least 5 years of industry experience, or 3 years with a PhD, in Data Science, Machine Learning, or Image Analysis.
- Demonstrate a working command of reproducibility and agreement statistics alongside medical imaging expertise and understanding of image data structures.
- Possess expert proficiency in Python and statistical data analysis ecosystems such as pandas, scipy, statsmodels, seaborn, and matplotlib.
- Have hands-on experience developing or fine-tuning deep learning algorithms using PyTorch for computer vision tasks applied to medical images.
- Show demonstrated proficiency using modern agentic tools and large language models as a daily force multiplier to accelerate software development.
- Demonstrate an exceptional ability to distill complex, multi-dimensional data analyses into clear, strategic insights for cross-functional stakeholders.
Nice to have
- Published research specifically related to domain generalization, image harmonization, or out-of-distribution detection in medical imaging.
- Experience working with large multi-vendor CT imaging datasets with an understanding of acquisition and reconstruction protocols.
- Proven track-record in diagnosing data- and annotation-related algorithm performance gaps and designing data-driven solutions.
- Experience with large-scale data querying and cloud storage like AWS and SQL.
- Experience developing Software as a Medical Device products, contributing to regulatory filings, and using biostats to support FDA submissions.
Benefits
- Bonus
- Equity
Skills and tech stack
Questions candidates ask
What is the salary range for this role?
A reasonable estimate of the base salary compensation range is $170,000 to $240,000, along with bonus and equity.
Where is the position located?
The role is located in San Francisco, California, and features hybrid work arrangements.
What tools and technologies will I be using?
You will use Python and statistical data ecosystems like pandas, scipy, statsmodels, seaborn, and matplotlib, alongside PyTorch and modern agentic tools such as GitHub Copilot, Gemini, and Claude.
What level of experience is required?
The company requires 5+ years, or 3+ years with a PhD, of industry experience in Data Science, Machine Learning, or Image Analysis.