Sophia Hart Email & Phone Number

Burlingame, California, Us

• Developed a linear regression model (Python) to optimize resource allocation across 250+ clinics, increasing service availability by 25%.• Led a project to refine provider quality measures using Bayesian statistics in SAS, reducing clinician maintenance time by 40% and driving policy changes that impacted over 65 million Medicare beneficiaries.• Managed a $1B Medicare payment monitoring project, analyzing data to identify $50M in potential savings, and communicated actionable insights to Medicare external stakeholders, influencing federal healthcare policy.• Conducted research on advanced Medicare technologies, resulting in a policy change that increased reimbursement rates from 65% to 80%, benefiting over 1 million patients nationwide.• Collaborated cross-functionally with stakeholders to ensure successful project deliverables, effectively communicating complex findings to both technical and non-technical audiences.

• Strategized and prototyped a Large Language Model (LLM) application for EdTech using ChatGPT, achieving production cost improvements of $0.02 per resume and scaling the solution to support 10,000+ B2B clients.• Led the implementation of the Software Development Life Cycle (SDLC) for a conversational resume application, ensuring efficient development and successful deployment.

Berkeley, Ca, Us

• Health Equity: Predicted total days to diagnose metastatic TNBC breast cancer using real-world data to identify patients under inequitable diagnosis using Python, TensorFlow-Keras, and kNNImputer. Ranked top 25% of 1,600 WiDS Datathon participants.• Breast Cancer Classification: Predicted if a patient’s breast cancer is benign or malignant from histopathology images.• COVID-19 and Race: Preprocessed data and created visualization of COVID-19 cases and death impact by race, and their correlation with income.• Wildfire Detection using Computer Vision: Developed and deployed a customized Xception pretrained model on AWS SageMaker. Potentially save first responder labor costs by 80% and wildfire detection time by 90%.• Climate Change: Developed forecast using XGBoost and LightGBM models. Ranked top 20% of 2,000 WiDS Datathon participants.• Bank Loan Risk: Implemented machine learning models to assess customer loan repayment likelihood using R.

San Francisco, California, Us

• Optimized differentiation of stem cells into neuron cells and analyzed biomarkers of these transplanted cells in mice models, leading to a 60% reduction in adverse side effects and enhancing treatment effectiveness for 200+ preclinical subjects.

Berkeley, Ca, Us

• Modeled cell division signaling pathway and its oscillating dynamics (time-space) on the position of E. coli cell division.• Developed a novel mathematical simulation of dynamics inside a cell using C, achieving 80% reduction in computation time.

Albuquerque, Nm, Us

• Developed a new biomedical device microfabrication material, whose oxidation-reduction window was the widest on earth.

Santa Clara, California, Us

• Designed a new integrated circuit, expanding the programmable range by 35% in testing defective CPU cache memory cells.

Certificate Program, Data Science

M.Eng., Biomedical Engineering

M.Sc., Electrical And Computer Engineering

B.Sc., Electrical Engineering And Computer Science; Physics Minor

M.Sc., Quantitative And Systems Biology