data scientist for clinical AI

The company builds a clinical AI platform that combines LLM orchestration, healthcare data, clinical capabilities, and AI safety mechanisms to deliver personalized healthcare experiences. Its Data Science team develops and evaluates the intelligence behind the platform.

• Own the analytical side of ML/AI evaluation and monitoring, including drift thresholds, distributional analysis, and quality metrics
• Develop statistical methods for safety-weighted quality scoring and confidence calibration
• Compare AI and model outputs against clinical ground truth and evaluate performance across populations and scenarios
• Perform population percentile, distribution, and outlier analysis to support health scoring capabilities
• Evaluate retrieval quality, embeddings, and vector search performance as RAG capabilities evolve
• Analyze usage patterns, overlap, deduplication signals, and classification accuracy across clinical AI capabilities
• Partner with Data Scientists and ML Platform Engineers to turn analytical methods into repeatable production evaluation workflows
• Investigate unexpected model or system behavior and identify statistically meaningful changes or degradation
• Use Python and AI-assisted development tools to accelerate analysis while maintaining technical validation and engineering quality
• Document evaluation methodology, assumptions, results, and recommendations

• Strong foundation in statistics and Machine Learning, including distributions, probability, model behavior, and evaluation methodology
• Strong Python skills and ability to write clean, maintainable analytical and production-oriented code
• Hands-on experience designing and performing statistical analyses on real-world datasets
• Solid understanding of model evaluation, confidence calibration, statistical testing, and performance metrics
• Experience analyzing data distributions, identifying outliers, and defining thresholds and quality indicators
• Understanding of data drift, model degradation, and distributional monitoring
• Ability to design evaluation approaches when ground truth is incomplete, noisy, or domain-specific
• Good understanding of Machine Learning workflows and the relationship between data, models, evaluation, and production systems
• Ability to understand application logic, system behavior, and data flow beyond isolated analytical notebooks
• Strong analytical and problem-solving skills, including independently scoping ambiguous problems and structuring analyses
• Ability to communicate methodology, assumptions, findings, and limitations to technical and non-technical stakeholders
• Strong ownership mindset and ability to collaborate effectively within a small cross-functional team
• Professional working proficiency in English
• Будет плюсом: Generative AI, LLM evaluation, or AI safety/quality assessment; embeddings, vector search, retrieval systems, or RAG pipelines; Databricks, Spark, or other large-scale/batch data processing technologies; AI coding agents or assistants, with critical validation of their output; clinical or healthcare data; FHIR or HL7; Healthcare, Digital Health, or another regulated industry; production ML monitoring, experimentation, or evaluation infrastructure

• Условий в вакансии нет