Bailey is a clinical AI system that orchestrates LLMs, manages health safety evaluation, and integrates clinical data to help people navigate their health. You will join the work that turns safety and reliability into measurable analytics, supporting teams as Bailey’s drift detection, scoring, and retrieval evaluations come online.
This Data Scientist role is the analytical half of the clinical AI infrastructure. You will own the statistical work behind drift detection and distributional monitoring, safety-weighted quality scoring, confidence calibration, and ground-truth comparisons using clinical data stored in FHIR. You will also deepen health scoring analysis in the Databricks batching pipeline, and support RAG and clinical skills library evaluation.
What you’ll do
- Drift detection and monitoring: define drift thresholds and own the analytical side of drift detection and distributional monitoring.
- Safety-weighted quality scoring: perform the statistical work behind safety-weighted quality scoring.
- Confidence calibration: own confidence calibration analysis.
- Ground-truth comparisons with FHIR: compare model outputs against ground truth in FHIR data.
- Health scoring depth: support population percentile and distribution work, including outlier handling, in the Databricks batching pipeline for health scoring analysis.
- RAG and embeddings evaluation: support retrieval quality analysis and embedding evaluation for the team’s move toward embeddings and vector stores.
- Clinical skills library analysis: analyze usage frequency, identify overlap and dedup signals, and evaluate risk-tier classification accuracy for the clinical skills library.
You will work across four currently-queued workstreams. You will not be handed all four on day one; you should expect to ramp in through one or two areas with close pairing with a senior data scientist, then take on more as you build context.
What you bring
- Real statistics and machine learning fundamentals already in hand.
- Strong Python ability, shown through code you wrote and can explain line-by-line.
- Understanding of systems and control flow to effectively direct a coding agent when you use one.
- Independence in ambiguous areas, including comfort scoping your own analysis and knowing when to ask for help.
- Ambition and ownership, treating knowledge gaps as things to close and feeling accountable for outcomes.
Helpful experience
- Experience directing a coding agent as part of your workflow (nice to have).
- Exposure to clinical data standards such as FHIR and HL7 or health information systems.
- Coursework or project experience with embeddings, vector search, or retrieval evaluation.
- Familiarity with Databricks, Spark, or batch-processing pipelines.
- Healthcare or other regulated-industry experience.
Team and tech
You’ll join a small data science team (Sean, Kenan, Zack) building Bailey’s AI capabilities, including models, integrations, clinical skills, and orchestration. A new ML/Platform Engineer will own production infrastructure alongside this role, and you will receive direct senior mentorship with real ownership over your analysis.
Technologies: Python, Databricks, Spark, FHIR, HL7
Compensation and benefits
- Pay range: USD 110,000 - 140,000 per year
- Stock options
- Benefits
- Incentive pay for eligible roles
- Competitive total rewards package
