merged into Voltrondata

Search Infra:

Knowledge Graph and Location understanding

Growth & Traffic:

1) Team ML roadmaps and priorities

2) Design ML foundations: Feature engineering, data preprocess, training, offline evaluation framework, production monitoring/feedback.

3) Built user intent understanding, audience segmentation, targeting and content optimization machine learning models.

Homes Central Data Team:

1) lead company-wide data initiatives, providing support to decentralized data teams throughout the company.

2) Design and implemented visitor-user-device/cookie graph based on the linkage of user-device, user-cookie, and cookie-device.

3) Offline bots traffic detection.

Risk Machine Learning Infra and data pipeline using spark/scala, java, kafka, hdfs/hive, and python

1. Anti-abuse projects

2. Account Take Over (ATO)

3. Payment Risk

Deliver end-to-end statistical analysis, predictive modeling, and machine learning modeling to business owners, including communication with business users to understand business problem 2) Define model objective function and model segmentation based on behavior difference and business requirement 3) Data query/generation, pre-processing, feature engineering, and model development using Hive, Spark, and Python. 4) Model validation, implementation, and performance and stability monitoring.

• Lead eBay global behavior model, which provides a FICO-Like score, ranks order all eBay seller with respect to negative events( financial loss and customer experience). It has 9 segmentations based on behavior difference and business requirement. The score has many business use case, such as loss based recoup, limit assign, and user ratings.

• Developed Queue Priority Optimization models which provides a rank ordered referral/leads for agent to review with respect to loss and customer friction.

• Lead UK Financial risk model.

• Lead internal new feature generation, created 1000+ new internal features, select top 20 variables pushed into production, considering information and variation.

• UAT test, model implementation validation, model performance and stability monitoring