Job Description
The Division of Data-Driven and Digital Medicine (D3M) is recruiting an early- to mid-career Machine Learning / AI Analyst to design, build, and evaluate solutions with a primary emphasis on Natural Language Processing (NLP) and multimodal AI, including multi-omics. Beyond clinical research and translation, the role includes developing internal decision-making and productivity tools for the Department of Medicine. You will collaborate with clinicians, scientists, and operations partners to turn clinical narratives, structured data, imaging, waveforms, and multi-omics into trustworthy models and user-friendly tools.
Responsibilities
• Lead NLP and multimodal ML efforts across text (clinical notes), tabular EHR, imaging, biosignals, and multi-omics to solve high-impact clinical and operational problems.
• Prototype and iterate internal decision-support and productivity tools (e.g., workflow triage, quality improvement insights, operational dashboards).
• Build robust data pipelines and features; ensure data integrity, lineage, and reproducibility.
• Train, fine-tune, and evaluate models (traditional ML, deep learning, and LLM-based approaches, including retrieval-augmented generation).
• Partner with clinical and operations leaders to frame problems, define success criteria, and run pilots that demonstrate measurable value.
• Operationalize models with MLOps best practices (versioning, CI/CD, monitoring, governance) and documentation for safe, responsible use.
• Follow privacy, security, and compliance requirements (e.g., HIPAA) and contribute to model risk management and bias/impact assessments.
• Communicate findings to technical and non-technical audiences through clear write-ups, visualizations, and presentations.
Qualifications
Minimum Qualifications
• Bachelor's degree in Computer Science, Biomedical/Clinical Informatics, Data Science, Statistics, Engineering, or related field (Master's preferred).
• 2+ years (industry, health system, or academic) working with ML/NLP using Python and/or R; strong SQL for data wrangling.
• Hands-on experience with modern ML/NLP (scikit-learn, PyTorch/TensorFlow; spaCy/Hugging Face), experiment tracking, and reproducible workflows.
• Ability to translate clinical/operational problems into analytical solutions and to communicate results to mixed audiences.
• Curiosity, product mindset, and commitment to responsible AI in healthcare.
Preferred Qualifications (emphasis areas)
• Deep experience in NLP and LLMs (prompting, fine-tuning, evaluation) and RAG over clinical knowledge bases.
• Multimodal learning across text, tabular, imaging, biosignals, and multi-omics.
• Experience integrating or analyzing multi-omics modalities (e.g., genomics, transcriptomics, proteomics, metabolomics) and linking them to clinical outcomes.
• Experience working with EHR data and standards (e.g., OMOP).
• MLOps tooling (MLflow, Weights & Biases), containerization/orchestration (Docker, Kubernetes), and cloud platforms.
• Practical understanding of model governance, fairness, and human-in-the-loop evaluation in healthcare.
• Track record delivering prototypes or products used by clinicians/researchers; publications or open-source contributions a plus.
Work Arrangement: Position is based in New York, NY (Icahn School of Medicine at Mount Sinai). Hybrid flexibility may be available per departmental policy.
