Data Scientist

Di (Grace) Wang

+1-734-***-**** ******.*****@*****.*** linkedin.com/in/di-wang-grace

SUMMARY

Accomplished data scientist with 2 YOE in data analytics. Proficient in leveraging advanced machine learning models, including Neural Networks, Deep Learning, and Gradient Boosting, to drive business success. Seeking a full-time role to contribute to data-driven decision-making. EDUCATION

Cornell University Ithaca, NY 08/2022 - 05/2023

Master of Engineering, Operations Research and Information Engineering, Data Analytics, GPA:3.5 University of Michigan Ann Arbor, MI 08/2020 - 04/2022 Bachelor of Science, Major: Mathematics, Minor: Computer Science, University Honors, GPA:3.7 SKILLS

Programming: Python(Scikit-Learn, TensorFlow, Keras, Matplotlib), SQL, R, C++, MATLAB, Java, Javascript Databases: Hive SQL, Pyspark SQL, SQL Server, MySQL, SQLite, MongoDB, Oracle Database Tools & Frameworks: PyTorch, PySpark, Tableau, Power BI, MLflow, Optuna, Jupyter, Visual Studio Code DATA ANALYSIS EXPERIENCE

Axtria Leading data analytics and cloud software company in the life sciences industry Data Scientist(Decision Science Associate), Berkeley Heights, NJ 07/2023 - Present

●TopicRecommendationSystemEnhancement:

● Addressed the limitations of the existing XGBoost model for time series data by developing a topic recommendation model using RNNs (LSTM, GRU) in Keras-TensorFlow, which improved prediction accuracy, achieving an 80% recall rate and 75% precision, and increased engagement rate by 8%.

● Evaluated the model in PySpark for optimized scalability and efficiency compared to local setups, resulting in a 50% reduction in training time, and conducted SHAP analyses for feature assessments.

● Preprocessed diverse data types (patient data, insurance claims, multi-channel interactions), performed feature engineering and exploratory data analysis, created visualizations, and prepared data for modeling.

●MLOpsPipelineDevelopment:

● Identified inefficiencies in manual ML workflows and communication gaps, developed an MLOps pipeline in Jupyter Notebook integrating feature engineering, model training, and monitoring workflows.

● Leveraged Random Forest for modeling, Optuna for hyperparameter optimization, and MLflow for lifecycle management, reducing deployment time by 30% and improving performance consistency by 20%.

●Non-personal Promotion Optimization:

● Built a robust channel attribution model using XGBoost that leveraged multi-channel touchpoints(calls, display ads, emails, etc.) to quantify the impact of display impressions on total prescriptions.

● Devised a strategic plan for display banner marketing, optimizing promotion levels for each healthcare provider using model predictions, resulting in a 15% increase in prescriptions and a 10% revenue growth.

● Communicated complex analytical findings effectively to technical and non-technical stakeholders. Cayuga Health Partner(CHP) Largest healthcare network optimizing patient care in Finger Lakes region Data Analyst, Ithaca, NY 12/2022 - 05/2023

●Led a capstone project in a team of three to enhance network integrity for CHP healthcare network, leveraging HIVE SQL, Python, and Excel to analyze and manipulate a dataset of over 6 million insurance claims.

●Collaborated with cross-functional teams to ensure seamless integration of data insights into operational workflows, recognized for exceptional teamwork and rated as a top performer.

●Developed and presented strategies to optimize patient retention to CHP stakeholders, resulting in a potential 10% increase in retention rates and enhanced operational efficiency. Leadbank Technology Leading Fintech company in Shanghai Quantitative Research Intern, Shanghai, China 08/2021 - 11/2021

●Tested convergence of 100 stocks with a 500 stock index using Python to perform RankIC and Yield Spread, resulting in a 12% improvement in predictive accuracy for stock movements.

●Designed a profit-enhancing strategy based on stock index deviation behavior and Fréchet distance, validated using the Roll model, resulting in an 8% improvement in portfolio returns over three months.

●Authoredacomprehensivereportonthestrategy'sperformanceandpublisheditonthecompany’swebsite.

Contact this candidate