Sign in

Data Scientist with a background in UX design.

San Francisco, CA
August 01, 2020

Contact this candidate


Stephanie Yeoju Jung San Francisco, CA 415-***-****

Data Scientist with a background in UX design and a passion for Machine Learning. Strong experience in communicating with non-technical stakeholders and working in a fast-paced startup environment. PROFESSIONAL EXPERIENCE

Data Science Intern, Virgo Surgical Video Solutions, Inc., San Francisco Oct 2019 – Present

● Increased frame-level accuracy of endoscopy video processing algorithm used to pinpoint procedure start / end times from 94% to 97.5%, which improved user experience by reducing false negatives.

● Built a video classification system for identifying endoscopy procedure type with 96% frame-level accuracy and 94.6% video-level accuracy, which enhanced user experience by enabling automatic video tagging for search and filtering.

● Acquired, preprocessed, and annotated over 120 hours of medical video footage.

● Helped design database schema and data pipeline for storing and processing labeled videos. Tech Stack:Python, PyTorch, TensorFlow, OpenCV, GCP, Google AutoML, SQL, FFmpeg. UX/UI Designer, ExInno Co., Ltd, Seoul July 2015 – Aug 2016

● Designed UX/UI for core mobile and web app for Hotel Shilla’s enterprise content management systems.

● Communicated with non-technical stakeholders regarding product requirements and delivered product results. Marketer/UX Designer, Withus, Seoul Feb 2015 – June 2015

● Devised go-to-market strategy for initial e-commerce platform. SKILLS

Programming: Python (Pandas, NumPy, Scikit-learn, SciPy, SpaCy, NLTK, Flask), Database (PostgreSQL), R, ggplot. ML/DL: PyTorch, TensorFlow, Keras, FastAI, OpenCV. Distributed Computing: Apache Spark (PySpark, Spark ML), Hadoop. Tools: Amazon Web Services (S3, EC2, EMR, EB, RDS), Kubernetes, Google Cloud Platform, Tableau, Google AutoML, FFmpeg, Google Analytics, Git, Docker.

Modeling: CNN, RNN (LSTM, GRU), Random Forest, XGBoost, Recommender System, Linear Models. EDUCATION

M.S., Data Science, University of San Francisco June 2020 B.S., Information Statistics, Korea National Open University Feb 2019 B.F.A., Visual Communication Design, Kookmin University Feb 2015 PROJECTS

Federated Learning for Smart Mobile Keyboards Current

● Trained a 2-layer weight-tied GRU language model on Wikitext-2 with a test perplexity of 100.

● Reduced training time by 3x through the use of adaptive softmax and layer normalization.

● Currently working on training a language model on Wikitext-103 in a federated setting. Tech Stack: Python, PyTorch, Docker, AWS EC2.

Sparkle: A Multi-platform Medication Adherence Monitoring System Github: May 2020

● Developed integrated iOS, watchOS, and web apps ( that provide patients with automated prescription schedules and medication reminders, and doctors with a dashboard summary of patient adherence rates.

● Used XGBoost classifier to verify medication intakes to ensure accurate adherence statistics (0.93 F1-score). Tech Stack: Python, PyTorch, Flask, Docker, AWS Elastic Beanstalk, Amazon RDS (postgres). Automating the Refillment of Prescription Medication with Apache Spark Jan 2020

● Collected 680 recordings of medication intake from 17 individuals using an AppleWatch (0.7GB total data).

● Used distributed computing to process sensor data and build a gradient boosted classifier that detects low medication counts in prescription bottles in order to automate orders of refillment medication (0.80 F1-score).

● Paper to appear at the 42nd IEEE Engineering in Medicine and Biology Society Conference (EMBC 2020).

● Tech Stack: Apache Spark, Spark ML, Python, Amazon EMR.

Contact this candidate