Candidates About

Danny  Zhang 

Data Scientist

Willing to relocate: Anywhere

• Above 5+ years of experience in Machine Learning, Deep Learning with large datasets of Structured and Unstructured data, Data Validation, Predictive modeling, Data Visualization.

• Extensive experience in Text Analytics, developing different Statistical Machine Learning, Data Mining solutions to various business problems and generating data visualizations with Python, Pig, Hive, Spark, TensorFlow, Keras

• Hands on experience in implementing LDA, Naïve Bayes, Random Forests, Decision Trees, Linear and Logistic Regression, SVM, Clustering, neural networks, Principle Component Analysis and Recommender Systems.

• Proficient in Statistical Modeling and Machine Learning techniques (Linear, Logistics, Decision Trees, Random Forest, SVM, K-Nearest Neighbors, Bayesian, XG Boost, LightGBM) in Forecasting/ Predictive Analytics, Segmentation methodologies, Regression based models, Hypothesis testing, Factor analysis/ PCA, Ensembles.

• Generated data visualizations to model data into human-readable form with Matplotlib, Seaborn

• Worked and extracted data from various database sources like Oracle, SQL Server

Experience:

Data Scientist

Wells Fargo, NJ                                                                                                        Oct 2015 to Present

Description: Wells Fargo & Company is an American international banking and financial services holding company. It is the world's second-largest bank by market capitalization and the third largest bank in the U.S. by assets. It is specializing in credit cards, home loans, auto loans, banking and savings products. The purpose of this project was to fight against credit card fraud. My team mainly focused on rebuilding credit card fraud detection model, monitoring the model in production, taking action if model performance degrades and working closely with business team to onboard new model. 
Responsibilities: 
• Communicated and coordinated with other departments to collection business requirement 
• Worked on miss value imputation, outliers identification with statistical methodologies using Pandas, Numpy 
• Participated in features engineering such as feature creating, feature scaling and One-1/ot encoding with Scikit-learn 
• Tackled highly imbalanced Fraud dataset using under sampling with ensemble methods, oversampling and cost sensitive algorithms 
• Improved fraud prediction performance by using random forest and gradient boosting for feature selection with Python Scikit-learn 
• Implemented machine learning model (logistic regression, XGboost) with Python Scikit- learn 
• Optimized algorithm with stochastic gradient descent algorithm Fine-tuned the algorithm parameter with manual tuning and automated tuning such as Bayesian Optimization 
• Validated and select models using k-fold cross validation, confusion matrices and worked on optimizing models for high recall rate 
• Implemented Ensemble Models with majority votes to enhance the efficiency and performance 
• Designed rich data visualizations with Tableau 9.4 
Environments: Python (scikit-learn, pandas, Numpy), Machine Learning (logistic regression, XGboost), Gradient Descent algorithm, Bayesian optimization, Tableau

Data Scientist

Starcom Mediavest Group  New York, NY                                                         June 2012 to Oct  2015

Starcom Mediavest Group (SMG) is among the largest full-service media networks in the world, which emphasizes the power of data and technology to capture emotion, enable great creativity and deliver connected, valuable human experiences. 
This project is developed to help client evaluate brand health, redefine market positioning, and seek the best possible strategies for advertisement investment. Our team is responsible for data modeling using varies advanced machine learning algorithms and for visualizing and reporting the results in support of strategic decision making. 
 
Responsibilities: 
• Communicated and coordinated with other departments to collected client business requirements 
• Conducted data exploratory analysis on survey data to learn customer feedback and response 
• Utilized survey data to evaluate brand health, understand customer path-to-purchase life cycle and sharpen the product positioning with statistical methodologies such as hypothesis testing for both single and multiple answer questions using Python 
• Identified competing brands based on survey data with correspondence analysis using Python 
• Compared POEM (paid, owned, earned media) performance concerning various measurement of awareness with Bayesian network and SEM (Structural Equation Model) using Python 
• Estimated the impact of TV advertisement in order to evaluate ROI with Kalman filter using Python 
• Built and optimized TV program rating prediction models with machine learning algorithm such as lasso regression and random forest using Python 
• Cooperated with tech department and set up an automated system to select most effective marketing channels 
• Validated and selected models using K-fold cross validation methods, error metrics and worked on optimizing models for higher accuracy 
• Created data visualization using Tableau, , Python matplotlib,  Seaborn,MS Visio and PowerPoint 
• Reported weekly progresses and presented final results to partners 

Skill:

Machine Learning (5 years), Deep Learning (2 Years), Python (5 years), SQL (65years), Spark (2 years), Hive(3 Year), Pig(3 Year), Keras(1 Year), TensorFlow(1 Year)

Education:

M.S in Statistics

Drexel University, Philadelphia, PA                                                                                            June,2012

B.S in Statistics

Central South University, Changsha, China                                                                              June,2002