Wanmei Ou

Ontada is the technology and data business unit within the McKesson Corporation. Ontada develops and maintains the flagship iKnowMed electronic health records (EHR) and clinical decision support (CDS) software as SaaS solutions for oncology practices. Ontada also supports clinical trial operation and real-world evidence generation through a combination of software, data, and service offerings. AI Center of Excellence (CoE) • Established the AI CoE to develop Natural Language Processing (NLP) and AI models to extract insights from unstructured documents and complex data patterns • Operationalized large language models (LLM), ChatGPT and Llama 2, to extract biomarkers, cancer staging and other clinical domains from 100+ million documents to accelerate chart abstraction process • Disseminated best practices among data science community in technology, people and process in cloud migration and compute cost optimization Data Analytics Platform Enabling Clinical Trials & Value-based Contracts • Oversaw development and delivery of individual product strategies and roadmaps of Ontada’s data analytics platform, balancing short- and long-term priorities against business needs • Partnered with the White House Cancer Moonshot program to accelerate data standards adoption for outcomes-based reimbursement • Conducted clinical trial patient matching algorithm on mCODE database with combined structured and NLP extracted data, on average improving matching by 4X • Established an AI-driven clinical review software product with human-in-loop to reduce medical abstraction effort by 50% • Completed migration of 200 TB of diverse data from on-premise and AWS to Azure/databricks with industry-standard lake-house architecture in six months • Built agile teams and orchestrate processes to ensure timely delivery aligned with business goals

VA’s enterprise data and analytics platform, Rockies, enables data-driven decision making across the enterprise. This digital transformation effort includes fully data life cycle: data import, harmonization, manipulation, analysis and visualization of large-scale and complex electronic health records, patient reported health data, and social benefit records from tens of millions of veterans, staff, and family members. Developed based on the Agile methodology, Rockies went live after four months of intense development. It currently supports five business use cases and hundreds of analysts, spanning three service models: full-service, Platform-as-a-Service and Data-Science-as-a-Service. VA’s SMART-on-FHIR clinical decision support (CDS) platform offers interoperable application development building blocks for VA’s decade-long VistA-to-Cerner EHR migration journey. Its first application COVID19-Patient-Manager launched within five months of development, and it is live at more than 20 sites to argument clinicians’ decisions in triaging and disposition at the emergency department. • Established an enterprise cloud-native data and analytics platform product portfolio to improve VA capability in data standardization, integration, and analysis, as well as seamless insight delivery to user workflows. • Prioritized multi-year portfolio roadmaps through close collaboration from nine program offices • Delivered three data products: medical procession ratio, natural language processing medical concept extract from medical notes, and patient object model • Established three service pillars based on different stake holders’ needs: end-to-end use case development partner, Platform-as-a-Service (PaaS) partner, and self-service partner • Developed technical and strategic sourcing blueprints to benefit VA from the best-of-breed product solutions while minimizing vendor lock-in • Managed five scrum teams of architects, data scientists, engineers, DevOps

• Led real-world data platform development leveraging existing health technology standards such as Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM). • Led the CMS AI Challenge submission with Amazon AWS team and ranked top 25 out of several hundred submissions. Work featured in the media. Work featured in the media. • Delivered self-service applications on real-world treatment duration for key oncology products • Established a model framework to design value-based contract to improve healthcare affordability. Work featured in the media. • Directed analysis of real-world data to discover new drug targets or new indications in order to accelerate R&D and to bring patients more effective treatment • Managed a team of data scientists and outcome researchers on observational study design and execution

• Established precision medicine informatics product portfolio from investment proposal through functional design, to product release, serving over ten healthcare and life science customers • Prioritized multi-year product development roadmaps and created go-to-market strategies • Collected and analyzed use case requirements across therapeutic areas and health sciences divisions, such as oncology, neurology, infectious disease, translational research, and clinical development • Advocated interoperability standards, such as HL7 FHIR and GA4GH, in both internal product design and external community activities • Influenced customers’ multiple-year projects which focused on integrating EHR, disease registry, EDC, biobank, LIMS, and genomic data • Defined product pricing for on-premise and SaaS deployment models with a matrix team • Organized scientific program for product user group meeting for customers and prospects

• Improved detection accuracy of pathological brain structures via computer-aided analysis • Conducted cognitive, memory, and motor sensory experiments using multimodal imaging technology: MRI, DTI, fMRI, EEG, MEG, and Optical imaging

Fixed-income structuring • Accelerated business cycle by assisting sales desk with pricing tools for structure products • Designed an interest rate-equity-commodity hybrid product for risk-adverse clients • Awarded 1st place in the foreign exchange portfolio championship among 5 teams and 30 interns

• Sped up processing by a factor of 100 for image alignment through tailored optimization methods • Integrated the new image alignment module into the commercial system for day-to-day operation

• Led product integration of wireless ultrasonic sensors used to economize fuel tank distribution

• Analyzed and simulated inheritance models based on recombination fraction in population genetics • Examined risk factors in kidney transplants using a neural network