Data scientist and neuroimaging researcher in the Brain-Gene-Development Lab.- Developed a comprehensive process for curating and analyzing imaging data leveraging clinician text reports- Facilitated conversations within cross-functional team to collaboratively build an organization process for aggregating large-scale unstructured data- Built a data-driven model of brain growth charts for pediatric MRI scans with limited pathology (Schabdach 2023)- Identified research goals for multiple projects based on research objectives- Managed and organized data and computational resources- Onboarded new lab members- Served as primary point of contact with honest broker on behalf of labPrimary Research Project: Scans with Limited Imaging Pathology.- Developed a process for curating a set of clinically acquired MRI scans consisting of reviewing radiology reports, coordinating with hospital departments to collect and transfer select deidentified data, filtering MR images based on metadata properties- Processed MRIs using both industry standard tools and novel deep learning based tools to extract volumetric imaging phenotypes- Analyzed relationships between imaging phenotypes, patient demographics, and reason for patient scan- Paper under review, preprint available on medRxiv as 'Brain growth charts of “clinical controls” for quantitative analysis of clinically acquired brain MRI'

Worked in a magnetic resonance imaging lab with radiologists, physicists, biomedical informaticists, statisticians, and clinical coordinators. Primary project focused on reducing the effects of motion in resting-state functional MRIs, identifying age-group level motion patterns using machine learning techniques, and identifying the effects of motion correction on desired signal.- Implemented 2 frameworks for image volume registration to reduce motion in resting-state functional magnetic resonance sequences- Demonstrated novel method produced results 5x better than those of the traditional method- Collaborated with clinicians and physicist to developed a lightweight functional MRI simulator- Demonstrated that novel registration framework can improve image quality better than traditionalAccomplishments- Student Paper Award Winner, American Medical Informatics Association 2020 Virtual Informatics Summit- Defended as dissertation March 31, 2020- 2 published papers

Integral team member for 3 projects related to sleep and using machine learning. Coauthored 8 invention disclosures, 1 published patent to date.SmartSleep (Demoed at CES 2019):- Analyzed data from over 1000 users to identify shared complaints associated with common sleep problems- Visualized data patterns using unsupervised ML (agglomerative clustering/heatmaps)- Built deep learning frameworks to predict a user's sleep problems based on their responses to a clinically-validated survey- Developed and presented a plan to incorporate and expand the ML abilities of the projectSmart Home Sleep Assistant:- Explored the smart home/virtual assistant domain and evaluated existing technologies- Identified software and human-computer interface requirements- Utilized Rasa and Actions on Google to build a sleep-related conversational user interface (CUI)- Presented a live demo of the CUI of to department leaders as well as an international team - Worked with department members to identify other applications of a CUI in the companyNatural Language Sleep Terminology:- Led this project on a team of 3 interns- Organized free-text responses from a survey of over 1000 users into general categories- Developed a set of contextual standards for specifying concepts associated with different phrases- Collaborated with 2 team members to manually annotate over 900 instances of free-text sleep complaints- Investigated natural language processing techniques to better differentiate between similar phrases- Built an ML model to classify the concepts present in a sample of free text- Presented the model to the department and modified the concept classification based on expert feedback- Incorporated the model into the Smart Home Sleep Assistant

Registration of Mammography and Ultrasonography for Improved Breast Cancer DetectionDeveloped and tested a prototype pipeline for registering mammography and ultrasonography images for improved breast cancer detection for high-risk patients. Tested multiple 3D deformable models and evaluated effectiveness in the pipeline. Built a pipeline for 3D-2D image projection, unimodal affine (rigid) and B-spline (deformable) registrations. Illustrated how existing prototype could be expanded for multimodal, multidimensional image registration.Content-based Mapping for Mobile DevicesIntegrated a knowledge base registrar and a doctrine-based military ontology with an Android application to create an intelligent communications system ideal for deployed soldiers in areas with inconsistent or no cell service. Developed an Android application to showcase the importance and flexibility of the military ontology. Modified the military ontology to clarify various relationships and remove duplicate entries. Assisted with a field demo of the software to military personnel.

Automatic Growth Percentile CalculationsImplemented a rule-based calculation server and incorporated JavaScript add-ons into data entry interface. Created automatic growth-percentile calculation server Incorporated server and other JavaScript add-ons to enhance the data-entry interface.Genotype-Phenotype ClusteringUsed existing clustering algorithms to cluster high dimensional data for use in later research. Clustered an ontology of annotated genes by their displayed phenotypes for use in later research Implemented visualizations for high dimensional data to demonstrate clustering results.