📍 Hybrid, Palo Alto or San Francisco (Hybrid – 2 days/week in office)

Reports to: Director Data Science + Analytics

Job Type: Full-time W2

About the Role

We are looking for a pragmatic, hands-on Data Scientist to build and own the quantitative foundation for our clinical care delivery, patient treatment plan refinement, and retention engines.

In our specialty care model, clinical health outcomes and business growth are directly linked: matching patients to their preferred, personalized treatment path leads to higher clinical efficacy, better satisfaction, and long-term retention. In this role, you won't just work with clean tabular data; you will bridge the gap between structured clinical metrics (lab values, symptom severity scores, dosage trajectories) and unstructured patient communications (visit transcripts, ambient scribe outputs, and intake survey dialogue).

You will sit at the intersection of Data Science, Clinical Operations, and Product, serving as a core builder who transforms messy, multi-modal patient data into actionable clinical decision support and personalized care pathways.

What You’ll Do

Data Foundation & Pipeline Development

Clean & Transform Multi-Modal Clinical Data: Build reliable, well-structured modeling tables and feature stores by parsing, cleaning, and joining raw datasets including lab results, symptom severity surveys, medication history, and raw visit transcripts.
Establish Unstructured Data Workflows: Design robust data processing pipelines to convert unstructured clinical notes, orders and visit transcripts into structured, queryable relational tables.

Transcript Mining & Feature Extraction

Extract Patient Intent & Preference: Build lightweight NLP and LLM-based pipelines (using entity extraction, sentiment analysis, and topic modeling) to convert ambient visit transcripts into structured features. And provide insights into patient preferences in treatment pathway, communication, and modality, etc.
Identify Friction Points: Process qualitative dialogue to automatically detect unaddressed obstacles or early side-effect reports before they lead to treatment discontinuation.

Lookalike Cohorts & Patient Treatment Planning

Lookalike Clustering Engine: Develop unsupervised segmentation models to cluster patients into "lookalike" cohorts based on baseline lab values, symptom profiles, and transcript-derived preferences.
Outcome & Trajectory Forecasting: Build predictive models that estimate symptom reduction, time-to-relief, and potential side-effect risks across alternative treatment branches (e.g., immediate HRT vs. conservative step-therapy).

Causal Inference & Retention Analytics

Quasi-Experimentation on Care Protocols: Apply observational study methods (Propensity Score Matching, Difference-in-Differences, Survival Analysis) to evaluate non-randomized care pathways and quantify counterfactual patient outcomes.
Drop-Off & Retention Optimization: Identify critical clinical touchpoints and dosage adjustment milestones that drive long-term program adherence, connecting clinical milestone achievement directly to patient retention and LTV.

Research & Cross-Functional Collaboration

Evidence Generation: Partner with the clinical research team to provide analytical support for peer-reviewed medical publications, white papers, and payer studies.
Clinical Operations Alignment: Translate analytical findings, treatment protocol insights, and patient drop-off trends into actionable workflow improvements for clinical operations and care teams.

What You Bring

Technical Skills

Programming & Querying: Advanced proficiency in SQL for data manipulation alongside strong Python skills for predictive modeling and data transformation.
Machine Learning & Modeling: Practical experience applying clustering (segmentation), regression/classification, and survival analysis (e.g., scikit-learn, statsmodels, lifelines).
NLP & Modern LLM Workflows: Hands-on experience with NLP techniques and LLM frameworks (e.g., transformers, Claude API, etc.) to parse, extract, and classify unstructured text data.