Sinhee Cho Email & Phone Number

South Korea

대한민국 서울

[ Recommendation system for T-membership benefit curation service ] - Designed a recommendation engine for T-membership - Developed a framework for generating candidate (user-benefit combinations) through users' geographical features and the benefit braches' physical locations - Built models for benefit curation service (benefit recommendation model, contents ranking model) - Operated a A/B test for each model to improve models' metrics (vs. baseline models) - Developed a dashboard to monitor KPIs of the task and model A/B test result[ Recommendation model for Telco. POCs ] - Built models to predict users likely to subscribe additional plans at T-world app./offline channels

[ General purpose user embeddings based on mobile usage ] - To save resource for handcrafted feature enginnering, generated user embedding vector for various machine learning applications - Based on sequential mobile usage logs, built log embedding model and summarized each customer's logs to fixed-length vector[ Customer segmentation based on heterogeneous data sources ] - Developed framework for handling and integrating user behavior logs from diverse data sources (more than 10 sources) - Built models for discovering topics in the logs and identifying users' short term interests (topic modeling + manual tagging) - The output was applied to target marketing and served as a source of feature for machine learning applications (especially for explainable AI)[ Predictive modeling for cellular network QoE ] - Built models to predict customers dissatisfied with cellular services - Performed root cause analysis for dissatisfaction (explained the model's output using treeSHAP) - Summarized personalized dissatisfaction score in different levels (dissatifaction score for LTE Cell, Building, area)[ Analytics on location data ] - Built geographical feature store from mobile usage logs - Participated in developing location intelligence service (designed features and performed data engineering)[ Predictive modeling for customer service support ] - Built models to classify customers' intents for visiting customer service center - Service menus were sorted based on the predicted intents in Interactive Voice Response)[ Data science & predictive modeling for telco. marketing ] - Performed various EDAs to support target marketing - Built models for target marketing (Smartphone, Internet Plans)

대한민국 서울 강남구

BI consulting group internship

석사, 지식서비스공학

학사, 경영과학