Jae Han

Designed the overall architecture for a machine learning project in   cybersecurity to mitigate credential stung attacks using Google   Dataflow, Pub-Sub, GCS, and AI Platform.    ● Lead efforts to obtain real-time log in event data for big data pipelines   using Apache Beam and Python    ● Translated models from Apache Spark to Apache Beam for use in   credential stuffing models   ● Create Splunk Dashboards for investigating hours saved by automation   and other SIEM products (Phantom, Proofpoint Trap)

Designed and implemented a randomized testing algorithm to mimic   real-time failures for virtual server offerings at SoftLayer.    ● Implemented wide-coverage test automation suites in Python, Php, to   minimize code failures in QA and Production.    ● Researched and implemented a neural network for binary classification to   predict VSI failures, based on popular customer configurations.   ● Leveraged data science tools such as Numpy, Pandas, Matplotlib,   Seaborn, for EDA and deep-dive analysis

Developed automation feeds and programs to support marketing   channels and Crate affiliates via Jenkins, Python, and Google AdWords.    ● Researched machine learning and leverage Google AutoML to detect   Crate&Barrel products in online blogs and posts;    ● Implemented deep-dive and exploratory data analysis for marketing   campaigns and customer data    ● Leverage AdWords scripts to maximize ROI and spending in Google   AdWords marketing campaigns.   ● A/B testing between different search campaigns, and text ads / copies for   use in optimizing paid search revenues in Google AdWords.