Manager, Applied AI, Advanced Informatics

Scorpion Therapeutics; Village of Tarrytown (NY); Hybrid; USD 180,000 - 280,000; Full time

Job summary

Scorpion Therapeutics is seeking a seasoned AI/ML role to frame clinical and business informatics challenges into clear ML problem statements. You will design and build end-to-end ML pipelines, run rigorous experiments, and collaborate across teams to deploy models into real workflows.

The role requires strong Python skills, experience with NLP or multimodal models, and familiarity with cloud ML, as well as health data and medical ontologies.

Qualifications

• Bachelor’s degree in CS/ML/Data Science/Biomedical Informatics/Statistics or related field.
• Master’s or Ph.D. with 4–6+ years applied AI/ML preferred.
• Grounding in supervised, unsupervised, and self-supervised learning; deep neural networks.
• Strong Python skills (scikit-learn, HuggingFace Transformers, pandas, NumPy).
• Experience designing/evaluating NLP or multimodal models, including LLM fine-tuning.
• Experiment tracking/reproducibility (MLflow, Weights & Biases, DVC).
• Familiarity with cloud ML (AWS SageMaker, GCP Vertex AI, Azure ML).
• Experience with health/life sciences data (EHR/EMR, claims, clinical notes, genomic, imaging).
• Familiarity with medical terminologies/ontologies (SNOMED CT, ICD-10/11, LOINC, RxNorm, OMOP CDM).
• Publications/open-source work in applied ML/NLP/computational biomedicine.

Responsibilities

• Frame clinical and business informatics challenges into well-scoped AI/ML problem statements with clear success criteria.
• Design and build ML pipelines (data ingestion, feature engineering, training, evaluation) alongside production engineering.
• Run experiments, benchmarks, and ablations to validate model performance and guide modeling decisions.
• Partner across clinical informatics, data engineering, and ML engineering to bring models into real workflows.
• Track advances in foundation models, LLMs, and retrieval-augmented generation, and apply to biomedical/health data.
• Document research findings and contribute to internal reports, publications, or conference presentations.
• Explain model behavior, limitations, and performance to technical and non-technical teams.

Skills

Python programming

ML fundamentals

NLP modeling

Experiment design

Cloud ML

Education

Bachelor's in CS/ML/Data Science/BI Informatics/Statistics

Master's/PhD preferred

Tools

PyTorch

TensorFlow

HuggingFace

MLflow

Weights & Biases

DVC

AWS SageMaker

Job description

Discover Your Role

• Frame clinical and business informatics challenges into well-scoped AI/ML problem statements with clear success criteria
• Design and build ML pipelines (data ingestion, feature engineering, training, evaluation) alongside production engineering
• Run experiments, benchmarks, and ablation studies to validate model performance and guide modeling decisions
• Partner across clinical informatics, data engineering, and ML engineering to bring models into real workflows
• Track advances in foundation models, LLMs, and retrieval-augmented generation, and apply them to biomedical/health data
• Document research findings and contribute to internal reports, publications, or conference presentations
• Explain model behavior, limitations, and performance to technical and non-technical teams

This Role Requires

• Bachelor’s in CS/ML/Data Science/Biomedical Informatics/Statistics or related; Master’s/Ph.D. strongly preferred with 4–6+ years applied AI/ML (Ph.D. research experience considered)
• Grounding in supervised, unsupervised, and self-supervised learning; deep neural networks; modern ML frameworks (PyTorch/TensorFlow or equivalent)
• Strong Python skills (scikit-learn, HuggingFace Transformers, pandas, NumPy)
• Experience designing/evaluating NLP or multimodal models, including LLM fine-tuning or prompt engineering
• Experiment tracking/versioning/reproducibility (MLflow, W&B, DVC, or similar)
• Familiarity with cloud ML (AWS SageMaker/GCP Vertex AI/Azure ML or equivalent)
• Experience with health/life sciences data (EHR/EMR, claims, clinical notes, genomic, imaging)
• Familiarity with medical terminologies/ontologies (SNOMED CT, ICD-10/11, LOINC, RxNorm, OMOP CDM)
• Published/open-source work in applied ML/NLP/computational biomedicine
• MLOps/CI/CD for ML or model observability in production
• Familiarity with federated learning, privacy-preserving ML, or regulated data use agreements