Evidence-oriented learning track for healthcare machine learning and data science across clinical framing, HL7 v2/FHIR/DICOM/OMOP data contracts, cohort design, calibration, target-trial thinking, multimodal workflows, cloud deployment, and AI governance.
Frame health AI around decision timing, measurable endpoints, workflow fit, and staged evidence rather than notebook metrics alone.
Compare HL7 v2, FHIR, DICOM, and OMOP as the messaging, resource, imaging, and analytics contracts behind healthcare ML.
Build cohorts, labels, features, and data-quality checks that preserve temporal realism, phenotype meaning, and site-shift awareness.
Frame prediction tasks, test temporal and external validity, calibrate risk, and report models transparently.
Trace DICOM-aware imaging AI from curation and labeling through reader studies, workflow integration, and monitoring.
Use notes, reports, and multimodal context for extraction, retrieval, summarization, and clinician-reviewed support tools.
Move from OMOP harmonization to target-trial-inspired observational design, diagnostics, and decision-ready real-world evidence.
Map healthcare ML systems onto AWS and GCP architectures, governed data stores, release gates, and monitoring loops.
Apply WHO, NIST, BMJ, and FDA guidance to documentation, change control, transparency, and postmarket oversight.
Validate your knowledge across clinical framing, standards, cohorts, calibration, imaging, NLP, deployment, and governance.