Senior Data Scientist

Probably Genetic•San Francisco, CA

1d•$180,000 - $230,000•Hybrid

About The Position

We are looking for a Senior Data Scientist who will own some of the most consequential diagnostic AI in rare disease: building, validating, and operationalizing the models that help us find and diagnose patients who have never had a name for their disease, powering the analytical rigor behind our testing programs, and shaping how we use data to make smarter product decisions.

Requirements

7+ years of experience in data science, machine learning engineering, or a closely related field
Strong Python proficiency and fluency across the core data science stack: pandas, NumPy, scikit-learn, PySpark, and SQL
Demonstrated end-to-end ML experience: you have taken models from problem definition through feature engineering, validation, deployment, and monitoring in a production environment
Experience with NLP techniques and applying language models to real-world problems
Comfort with prompt engineering and evaluating external AI API performance (e.g., OpenAI)
A track record of operating with high ownership in lean, fast-moving environments where you have had to build structure as much as execute within it
Strong analytical communication skills — you can translate complex model outputs and data findings into clear, actionable narratives for technical and non-technical audiences alike

Nice To Haves

Experience with Databricks or similar lakehouse/ML platform environments
Familiarity with synthetic data generation techniques
Domain knowledge in healthcare, rare disease, genomics, or clinical research
Experience with MLOps tooling and building observability infrastructure from scratch
Exposure to biopharma or insurance analytics use cases

Responsibilities

Own the end-to-end development, validation, and operationalization of PG's predictive diagnostic AI models — from feature engineering through production deployment – that power program eligibility decisions and clinical decisions for patients
Run prospective testing experiments: apply diagnostic models to undiagnosed patients, coordinate testing, and track outcomes to continuously improve model performance
Build and maintain PG's synthetic patient data pipeline, a critical deliverable for our research programs, and key input to our own model development lifecycle
Optimize our patient intake experience using NLP and multimodal data analysis to determine which questions to ask, in what order, to maximize data quality and conversion
Own API usage and cost optimization across PG's AI stack, including prompt engineering, model evaluation, and ongoing performance monitoring
Conduct ad hoc strategic analyses that inform product prioritization, causality assessment, and generate customer-facing program insights
Establish MLOps infrastructure: model monitoring, drift detection, API observability, and lightweight but durable operational processes
Have the freedom to conduct blue sky research initiatives aimed at creating value from our data
Work with Data Engineering to build a robust, scalable data foundation that supports all of the above