My research will be focused on multimodal AI to create, validate, and implement parsimonious methods to improve healthcare. Interpretability, explainability, reduction of bias, and patient safety will be foundational to this work to make it trustworthy for clinicians and patients.

Part-time ongoing position concurrent with various full-time roles.AI/ML research: Sepsis detection and management using probabilistic (Bayesian/Markov) models and Neural Nets.Teaching: Instructor and guest lecturer of graduate-level health data mining, machine learning, terminologies, ontologies, and standards courses.

Part-time ongoing position concurrent with various full-time roles.CDC NHSN Antibiotic Use: Led data standardization, designed, built, and implemented a centralized system to report antibiotic usage from 100+ VA hospitals to CDC in HL7 CDA format using the secure NHIN Direct protocol.CDC NHSN Antibiotic Resistance: Led data mining and analytics to find trends of antibiotic resistant bacterial infections and validated the CDC NHSN Antibiotic Resistance reporting protocol using nationwide VA data.Synthetic Data: Created synthetic datasets and validation methods for CDC NHSN Antibiotic Use and Resistance (AUR) protocols, which are used by CDC to validate vendor systems. Advised multiple vendors on behalf of CDC.

Incubated at The University of Chicago Booth School of Business Center for Applied Artificial Intelligence, Nightingale levels the playing field for academic clinical and computational researchers worldwide to have free access to multimodal clinical data for discovering new medical knowledge.Role: Strategic, Operational, Product, and Data Science lead of NightingaleFundraising: Obtained funding from NIH and private foundations to extend runway by 18 months. Reduced cloud costs and passed savings to end-users. Grew user community by 200%. Partnerships: Built and strengthened partnerships with Stanford Medicine, Providence Healthcare, Brigham and Women's Hospital, Emory University, Health Ministry of Tamil Nadu, India), and others for publishing deidentified data sets.Big Data: Published open-access medical data sets for AI research on including the world's largest open cancer biopsy data set. Created data specifications and quality requirements. Reviewed data and models for validity and bias.Data Science: Led data science team in building deep learning models for identifying cancer staging, tuberculosis, and fractures. Conducted 3 ML challenges on cancer detection and health equity with >75 teams worldwide.New Initiatives: Launched teaching labs with courses taught by MIT, UC Berkeley, NIH, University College London, and others. Launching institutional partnerships with multiple universities.Research: Evaluating large language (LLM)/foundational models for clinical validity, safety, and medical AI research.

Research Associate (WOC) at Salt Lake City VA Medical Center through my part time appointment at the University of Utah School of Medicine. Research primarily involves infectious diseases, antibiotic usage and drug resistant bacterial infections through the analysis of electronic health record data from the VA system.

Company: Truveta provides a research and data platform with deidentified data sets of 100 million patients from 30 US health systems.Role: Product/Informatics leadership of two multidisciplinary teams of forty engineers, informaticists, physicians, nurses, pharmacists, and data scientists.Product: Product lead for syntactic and semantic normalization of data from Epic Clarity and Cerner electronic health records, terminology standardization, natural language processing and machine learning. AI: Created product roadmap for data normalization and platform using human-in-the-loop artificial intelligence. Standards: Mentored engineering teams about HL7 FHIR, CQL, value sets, various standard terminologies, terminology mapping, OMOP/OHDSI, PCORnet, and implementation of standards. Research: Conducted and published research on COVID 19 vaccines, adverse events, and health equity.

Strategic and operational leadership of health care data strategy, data governance and analytics at 3M Health Information Systems, Inc. Served as interim Chief Data Officer of 3M HIS from 2017 to 2018 with similar responsibilities before promotion to current role.Responsibilities:- Companywide clinical data strategy, governance and acquisition of health care data for research and product development- Development and operation of secure centralized data and analytics platform- Development and validation of predictive models and technologies

Terminology and Ontology Research and Product Development, Strategic and tactical leadership of 3M HIS clinical data governance .- Research and design of 3M Healthcare Data Dictionary server software and terminology content- Design, validation and implementation of the 3M HDD as a FHIR-enabled multitenant cloud-hosted healthcare terminology server- Research and development of advanced terminology search engines and user interfaces for clinical terminology management. - Leading the clinical data strategy of 3M Health Information Systems and implementing a data governance framework- Various standards development activities including HL7 FHIR, Common Terminology services and other standards. - Curating and integrating terminology content (including SNOMED CT, LOINC, RxNorm and various flavors of ICD-9 and ICD-10 and some other terminologies) in the Healthcare Data Dictionary.

Externally funded research by 3M at the University of Utah. Healthcare Terminology and Ontology Research and Development.

My PhD research at the University of Utah School of Medicine was done at Intermountain Healthcare at the lab of Prof. Peter Haug, MD. My research on methods to predict future outcomes of patients as well as recommending interventions to improve future outcomes was performed using patient data from Intermountain Healthcare hsopitals. Real world patient data were used for early detection of sepsis and recommending insulin dosing for ICU patients for glycemic control.

Developed a patient tracking system for Cardio Thoracic Surgery, which was integrated with eChart, a home-grown web-based Electronic Medical Record system.Developed a bacteremia surveillance system (rule-based clinical decision support system) for the University of Utah Epidemiology Department by analyzing HL7 data streams.Evaluated methods to automatically map domestic vocabularies to UMLS.

Developed a research prototype for a Guideline Representation Language, Guideline Instance and a Guideline Execution Engine for web based, wizard-style implementation of Clinical Practice GuidelinesPerformed information modeling, and created XML schemas for a prototype Guideline Representation LanguageDeveloped a Guideline Execution engine in PHP utilizing XML parsers. Developed a prototype front-end for Wizard-style guideline implementation