Ricky (Shim-Young) Lee Email & Phone Number

Palo Alto, California, United States

- Developed and shipped ML models predicting key ad metrics (including ad revenue, install rate and event rate) that power AppDiscovery, a user acquisition service which drove over $300 million in quarterly revenue using BigQuery, Spark, and PyTorch.- Managed deployment and A/B testing of models developed; each improvement boosted advertiser spend and margin by up to 10%.- Created a gradient-based feature importance measurement tool using PyTorch Captum to guide feature and model… Show more - Developed and shipped ML models predicting key ad metrics (including ad revenue, install rate and event rate) that power AppDiscovery, a user acquisition service which drove over $300 million in quarterly revenue using BigQuery, Spark, and PyTorch.- Managed deployment and A/B testing of models developed; each improvement boosted advertiser spend and margin by up to 10%.- Created a gradient-based feature importance measurement tool using PyTorch Captum to guide feature and model architecture development for the first launch of deep learning models.- Implemented ML metric logging and monitored 1000+ models in production using Weights & Biases and Grafana. Show less

Palo Alto, California, United States

- Automated the aggregation of costs associated with fraudulent marketing channels using Airflow, which saved $500k+(~5% of spend) of monthly cost and replaced manual efforts with auto-generated reports.- Designed and implemented new signals for a fraud detection system that multiple teams of marketing analysts used daily for campaign optimization, using Pandas, Spark and MySQL.- Led technical communications with ad networks (Unity, Tapjoy) for refund negotiations, by explaining… Show more - Automated the aggregation of costs associated with fraudulent marketing channels using Airflow, which saved $500k+(~5% of spend) of monthly cost and replaced manual efforts with auto-generated reports.- Designed and implemented new signals for a fraud detection system that multiple teams of marketing analysts used daily for campaign optimization, using Pandas, Spark and MySQL.- Led technical communications with ad networks (Unity, Tapjoy) for refund negotiations, by explaining statistical methodologies used for fraud detection. Show less

Palo Alto, Ca

Graduate Data Science Internship, Ad Technology predictive modeling- Project: Deep semantic embedding for sponsored products- Researched and developed a deep learning based embedder for Sponsored products using TensorFlow, designed for ranking, filtering and other downstream tasks.- Implemented a multimodal (text + image) two-tower architecture model, trained on a data set of 10+ million user purchases.- Created visualization and retrieval demos for in-depth qualitative analysis… Show more Graduate Data Science Internship, Ad Technology predictive modeling- Project: Deep semantic embedding for sponsored products- Researched and developed a deep learning based embedder for Sponsored products using TensorFlow, designed for ranking, filtering and other downstream tasks.- Implemented a multimodal (text + image) two-tower architecture model, trained on a data set of 10+ million user purchases.- Created visualization and retrieval demos for in-depth qualitative analysis and final presentation. Show less

Gwanak-Gu, Seoul, Korea

Seoul National University Big Data Institute- Project: Analyzing real time soccer match data with a simple neural network (possession prediction)- Worked with ACM DEBS 2013 high-frequency sensor data generated from the sensors attached to players in a soccer match.- Implemented basic (multi-layer perceptron and convolutional neural network) deep learning algorithms with Tensorflow Also used many other classification algorithms to provide benchmarks using Python library… Show more Seoul National University Big Data Institute- Project: Analyzing real time soccer match data with a simple neural network (possession prediction)- Worked with ACM DEBS 2013 high-frequency sensor data generated from the sensors attached to players in a soccer match.- Implemented basic (multi-layer perceptron and convolutional neural network) deep learning algorithms with Tensorflow Also used many other classification algorithms to provide benchmarks using Python library scikit-learn.- Performed data cleansing, correction, manipulation as well as feature engineering extensively. Show less

Songpa-Gu, Seoul, Korea

Medical Imaging & Robotics Laboratory- Project: Computer aided diagnosis of breast cancer (analysis of 3D automated breast ultrasound images) - Implemented digital image processing and machine learning-based classification algorithms in MATLAB. - Assisted other researchers in translation, literature research and summary.- Experienced collaborative research in medical imaging and clinical science.

Master Of Science (M.S.), Statistics: Data Science

MS in Statistics/Computaional and Matematical Engineering (Track: Data Science)

Bachelor Of Science (B.S.), Applied Mathematics

Cum Laude