Data Scientist
Current- Pharmaceutical Process Yield Prediction and Production Optimization (Daewoong Pharmaceutical Co.) • Develop yield prediction model according to the changing outdoor temperature, humidity and the difference in parameters for each process• Recommend a combination of process parameters that optimizes yield according to a given temperature, humidity using a predictive model- AutoRL research for differentiating AccuTuning products• Derivation of development requirements for RL commercialization• Explore and define areas that can be automated in RL (Custom environment, agent, parameter)• Design AutoRL product pipeline• Develop web advertisement recommendation POC and review feasibility- Customer Churn Prediction in Vitality App (AIA Life Insurance Co.)•Comparative EDA of inactive and high-active user •Merge large app log, member information and activity data and present them as summary variables for predictive models•Develop LightGBM model and verify robust performance through cross-validation by monthly period •Analyze the contribution of Vitality app menu to customer loyalty with SHAP- Few-shot Model for Short-term Time-series Forecasting • Research on the latest technological trends for short-term time series forecasting (Self-supervised vs. Few-shot)• Adoption of few-shot method through case studies• Implemented robust few-shot model for various time-series data
