Prashant Kumar Email & Phone Number

San Juan, PR

San Juan, PR

• Use Statistical Knowledge for designing and analyzing A/B tests to drive KPI improvements, ensuring that we run the experiments with proper sample sizes and level of power.
• Providing insights and recommendations through advanced data analytics to business stakeholders and executive leadership based on data from various sources (backend, click stream/event data).
• Provide expertise, direction and oversight to product business intelligence and data engineering efforts to advance data infrastructure, governance and capabilities. Maintain and drive execution of product’s business.
• Wrote complex SQL queries to create multiple visual dashboard reports; these reports helped executives to make business decisions.
• Interact with and support a variety of teams and individuals within the organization to make data-driven business decisions aimed at growth/business optimization
• Train analysts in analytics best practices, as well as ensure continuous improvement of analytics methodologies

New York, New York, US

• Built the prepayment model for corporate forecasting using Neural network and GLM which will help the company predict and control profitability, and adjust future product prices.
• Built the charge off model to improve the collection strategies. Model predict the likelihood of an account going to Charge off state in next 12 months. Used LightGBM to build the model and saved ~$50 M annually
• Analyzed the 3.7M consumer credit bureau data to identify the high risk merchant for GreenSky
• Performed the relevance analysis on Merchant Lexis Nexis risk data and found the top risk features which can be useful to predict risk associated with the business (Eg. bankruptcies and business closure)

New York, New York, US

• Built model monitoring tool using python and tableau to evaluate all the production model efficiency over a period.
• Used PSI/CSI, KL Divergence, JS Divergence, Chi Square test, KS test and modern scikit learn metrics (Regression: RMSE, MAE, MAPE, MSE etc. Classification: Accuracy, Precision, Recall, F1 score, F Beta, AUC, ROC etc.).

New York, New York, US

• Developed a multi-media capable customer service Artificial Intelligence Chatbot using AWS Lex to explain the nuances of different types of plans to customers.
• The Chatbot has the potential to add 5% reduction in customers queries call initially, which would be like 7,500 calls a month, which translate to 7 FTE save. At 50k$ per FTE, GreenSky potentially looking for 350k$.

Atlanta, Georgia, US

• Worked on research project to identify licensing agreement between two companies using NLP and deep learning algorithm like CNN, RNN and LSTM.

Atlanta, Georgia, US

• Designed a credit score system for each broker by benchmarking Broker strengths and weaknesses using the Machine Learning model for Starr Insurance Companies.
• Employed K-means and MICE algorithm to perform null value imputations on raw data
• Implemented N-dimensional spherical projection bounding 95% of data for outlier detection.
• Created a mathematical model for getting a score of the brokers using constrained PCA algorithm, which reduces the underwriter speed by 30%

Dublin 2, IE

• Increased profits earned in the transportation sector for our client by 11% per quarter by building a machine learning model to predict the number of commuters on a given day for a specific region.
• Segmented marketing targets by commuter’s profile using K-mean & PCA. Results obtained helped client determine business strategies.
• Worked on the MYSQL Database server where real-time transactions generated from the transportation system in Canada are stored. Used to extract and analyze those transactions from the MYSQL server through SQL and sent.
• Created complex analytics dashboards using Tableau, extracting data from multiple data sources and then analyzing the data for pattern recognition in commuters’ profile and traveling time.
• Developed business logic modules to support loyalty and period passes for the current transit customers, increasing the number of commuters traveling by 10k per year.
• Conducted 4 knowledge transfer sessions(K.T) for 15 new team members, ensuring a smooth transition to project technologies

Master Of Science - Ms, Analytics

B.E, Electronics And Communications Engineering