Architecting predictive healthcare models and robust, HIPAA-compliant data pipelines to translate complex clinical ecosystems into life-saving, actionable insights.
A federated learning model deployed across 4 distinct healthcare networks allowing decentralized training on MRI scans without migrating sensitive patient data. Resulted in a generalized model for early-stage tumor detection with 94% AUC.
View Protocol
Built a custom middleware solution that standardizes inbound CCDA documents into FHIR R4 resources. This enabled seamless data exchange between legacy outpatient clinics and a modern centralized hospital Epic EMR system.
View Architecture
Developed an end-to-end real-time monitoring dashboard for ICU units. The engine analyzes vital signs and lab results every 15 minutes, alerting nursing staff of impending sepsis shock up to 6 hours before clinical manifestation.
View Metrics
Implemented a custom BERT-based Natural Language Processing pipeline to extract unstructured ICD-10 codes, medication dosages, and patient sentiments from raw physician narrative notes, automating billing workflows.
View Pipeline"Elias has a rare ability to translate complex clinical needs into highly efficient, scalable technical architectures. His work on our predictive models directly improved our ICU patient outcomes."
"I've worked with many data scientists, but Elias's deep understanding of healthcare compliance and FHIR standards sets him apart. He builds pipelines that are not just fast, but impenetrable."
"A true visionary in bioinformatics. Elias led our research team with agile methodologies previously unheard of in academic settings, delivering our oncology algorithms months ahead of schedule."
Currently available for freelance consulting, advisory roles, or full-time senior mission engagements in health tech.