Ethan Hsiao Email & Phone Number

San Mateo, California, United States

Palo Alto, California, United States

➢ Performed 100+ IR investigations with Splunk, automated and parellelized workflows in Torq, increasing visibility and security, saving 1k+ business hours, filtering out >50% of false positives➢ SOAR and JIRA automation with Slackbot integration expedited operational responses to threats, enhanced overall DART response by reducing manual intervention and potential errors➢ Migrated all hive templates/workflows from hive4 to hive5 with AWS, Cribl, Kubernetes, Python

Durham, North Carolina, United States

➢ Migrated 10 onboarding and security APIs from C# to Python for ECS PAM (Enterprise Cybersecurity Privileged Access Management) team; resolved bug fixes, added exhaustive unit tests (PyTest), & added automation, significantly improving code readability, scalability, and performance by >90%➢ Flask app deployed using Docker, tested using Postman, pipelined and productionalized using Jenkins, hosted on cloud using AWS; repository serving as model for numerous other migration projects

Pasadena, California, United States

➢ Developed python APIs to determine spatial coverage & revisit frequency for NASA satellites using GeoPandas, Bokeh, QGIS; works for any region-of-interest (spatial area) and temporal window➢ The APIs leveraged a JPL database of footprints from multiple satellites (e.g. Landsat and Sentinel-2), synced these footprints to the acquisition frequency, & performed relevant queries to determine the revisit frequency, coverage within a temporal window and spatial area, & study surface water, surface disturbance, and surface deformation products

San Mateo, California, United States

➢ Developed & pipelined ML model categorizing users’ complaints using PySpark, AWS, Jenkins➢ Model’s accuracy over 93% (BERT), tuned hyperparameters (batch size, learning rate, # epochs) ➢ BERT model currently used by Merchandising team to provide automated insights into NPS & user comments, driving business investments and more effective business strategy➢ AI model also utilized to analyze difference in distribution of complaints between brands, subscription programs, & root and subcategories of makeup➢ AI classifier highlighted frequency of repetitiveness complaints about brushes, leading Merchandising team to decrease number and frequency of brushes, improving customer experience➢ Demoed model at Tech Demos and All Hands Meeting (~500 people, including CEO & CTO)

Reston, Virginia, United States

➢ Built a humor detection model with an accuracy of over 95%, using a combination of naive rules, Random Forest Classifier, and FastText (uses Word2Vec)➢ Implemented a Flask application service deployed using Docker to train ML models, make predictions, and test trained models using k-folds cross validation➢ Presentation: https://www.youtube.com/watch?v=eiJ8as6wXRE➢ Github: https://github.com/ETH88/Humor_Detection

Cambridge, Massachusetts, United States

➢ Member of team creating ride-sharing self-driving bike (LiDAR), instrumental in testing vehicle➢ Improved autonomous bike simulation, refactored code, and created recharge/wait model, adding realism to model➢ Persuasive Electric Vehicle (PEV) project overview: www.media.mit.edu/projects/pev/overview

Bachelor Of Arts - Ba, Computer Science And Data Science

Activities and Societies: Machine Learning Developer at Launchpad (ML Consulting Club), Data Science Society, Theta Tau Engineering.

High School Diploma, Weighted Gpa: 4.89 / Unweighted Gpa: 3.98

Activities and Societies: Founder and leader of Computer Science and Economics club; starter on Basketball team; peer tutor for math.