Yunyi She

Designed and integrated a DeepSeek-powered preference-extraction workflow for a two-traveler recommendation service. Built versioned live and offline evaluation pipelines for ambiguity, corrections, provider failures, and prompt injection.

Delivered fleet-level database security reporting in Oracle Data Safe and led backend delivery for Database Vault integration. Improved assessment reliability by resolving scheduling, reporting, pagination, authorization, and concurrency defects.

Built structured EHR and clinical-text pipelines on MIMIC-IV to compare classical ML and ClinicalBERT-family models for ICU length-of-stay prediction. Achieved 0.805 AUROC using XGBoost with ICD-derived embeddings and first-authored a paper in Healthcare Technology Letters.

Developed full-stack resume-generation workflows for an AI-assisted product, migrating Python APIs from Flask to FastAPI and integrating them with React user flows. Secured backend APIs with Auth0 OAuth 2.0/OIDC, JWT authorization, role-based access control, and automated tests.

Developed Ruby on Rails APIs and React interfaces supporting automated billing, payment processing, user preferences, and report generation. Added 50+ RSpec and Selenium tests covering backend behavior and end-to-end workflows.

Master of Science in Computer Engineering with a specialization in Artificial Intelligence and Machine Learning, completed in June 2025.

Bachelor of Science in Mathematics and Computer Science, completed in March 2023.