Sarah Ki
Corona
Overview
3
3
years of professional experience
Work History
System Engineer - Reliability & Analytics
The Boeing Company
05.2024 - Current
- Analyzed large-scale 777 operator fleet datasets to evaluate reliability performance and deliver data-driven insights informing operational decisions.
- Built predictive regression and time-series models in Python to forecast climate-related impacts and support strategic planning.
- Automated data extraction, transformation, and analysis workflows using Python, reducing processing time and improving analytical efficiency.
- Developed automated data quality checks in Python to validate operator datasets, improve data integrity, and enable high-quality downstream analysis.
- Collaborated with customers and cross-functional stakeholders to ingest operational data, compare fleet performance against Boeing reliability recommendations, and translate insights into actionable recommendations.
Capstone: Transparent ICD-9 Coding Assistant
UC Berkeley
Berkeley
05.2024 - 12.2025
- Designed and built an end-to-end ICD-9 code recommendation system that automatically assigns top-k diagnosis codes from clinical discharge summaries (MIMIC-III), reducing manual medical coding effort while improving auditability.
- Implemented a hybrid retrieval–reranking pipeline that combines fine-tuned medical text embeddings (MedCPT) with large language model rerankers (Gemini Flash) to improve ICD-9 code relevance and ranking quality.
- Developed evidence span extraction to surface concise, human-readable clinical justifications for each predicted code, enabling transparency for medical coders and patients.
- Evaluated system performance using Precision@k, Recall@k, Micro-F1, and Macro-F1, beginning with a focused diabetes ICD case study and extending evaluation to the full ICD code set.
- Built an interactive Streamlit web application supporting coder workflows (review, accept/reject codes) and a patient-facing view for code transparency and trust.
System and Data Analyst - Flight Operations Support
The Boeing Company
11.2022 - 05.2024
- Created a Python tool to automate the BCS Service Request analysis for different operators
- Constructed operational and interactive report in Jira dashboard using MDX to query and visualize data on Flight Operations Support product quality escapes
- Analyzed and identified quality metrics (M3s) to improve Flight Operations Support products up to 80%
- Integrated application components and databases across computing platforms using Teradata to data mine real-time Flight Operations data
- Communicated business requirements and test results with offshore developers to manage projects
- Developed and executed tests to validate system functionality against specification
- Coordinated with Flight Operations teams to develop Python tools to automate workflow tracking processes with Jira REST API and Python scripts
- Performed research of process, applications, systems and data to support identification of functional requirements for application or system design
- Interpreted and translated application operational requirements into functional specifications
Education
Master of Science - Information and Data Science
University of California, Berkeley
Berkeley12-2025
Bachelor of Science - Bioengineering, Bioinformatics
University of California, San Diego
San Diego01.2021
Skills
- Programming: Python, SQL
- Machine learning: regression, classification, CNNs, time series
- NLP & LLM systems: embeddings, retrieval, and reranking, seq-to-seq models (T5, BART), LLM APIs (Gemini) for relevance scoring
- Evaluation & experimentation: Precision@k, Recall@k, Micro/Macro-F1, A/B testing
- Visualization & Communication: Tableau, Matplotlib, ggplot2
- Tools & Platforms: AWS, Docker, Git, Streamlit
Projects
- ICD-9 / ICD-10 Recommendation System: Built an end-to-end medical coding system using hybrid retrieval and LLM-based reranking to recommend relevant ICD codes from clinical notes, scaling from a diabetes-focused prototype to all ICD-9 codes and evaluating performance with Precision@k, Recall@k, and F1.
- Data Visualization & Policy Analytics: Developed interactive dashboards and visual narratives using Congressional bill data to analyze legislative activity across sessions, regimes, and policy areas, translating complex political data into accessible insights for non-technical audiences.
- Korean NLP & Cross-Lingual Processing: Researched and implemented Korean NLP models for grammatical error correction and cross-lingual retrieval, leveraging large-scale Korean text datasets and sequence-to-sequence architectures to address gaps in informal Korean language tooling.
Timeline
Capstone: Transparent ICD-9 Coding Assistant
UC Berkeley
05.2024 - 12.2025
System Engineer - Reliability & Analytics
The Boeing Company
05.2024 - Current
System and Data Analyst - Flight Operations Support
The Boeing Company
11.2022 - 05.2024
Master of Science - Information and Data Science
University of California, Berkeley
Bachelor of Science - Bioengineering, Bioinformatics
University of California, San Diego