S Dilip Kumar

Data Scientist Intern at Flip Robo Technologies & pursuing Post Graduate Diploma in Applied Statistics from Data Trained Institute(collaboration with IBM). Having an exhilarating experience of 6year. Proficient in preprocessing, analyzing large datasets, identifying interesting patterns, extracting valuable insights, building predictive models using data analytics & machine learning with some hands on Deep learning, CNN and Lexical Processing in NLP and model deployment.

K E Y D ATA S C I E N C E P RO J E C T S | MAR 2 0 2 1 - SEP 2 0 2 1

• Domain: Banking – Credit Card | Tech Stack : Python, Jupyter Notebook, Google Cloud AI Platform | Mar’20

• Objective: Building a credit card fraud detection model for a bank

• Solution: Did EDA, treated data imbalance using SMOTE &ADASYN, built 7 models&did Hyper parameter tuning for best performing models in each category

• Key Achievements: ADABOOST on SMOTE oversampling having AUC, recall & precision of 0.94, 0.90, 0.99 respectively

• Domain: Telecom | Tech Stack: Python, Jupyter Notebook | Nov '20

• Objective: Predicting the high value customers’ churn &identifying important attributes

• Solution: Designed 4 ML models via PCA, logistic regression, Decision Tree & Random Forest to predict the high value churners and determined the important variables

• Key Achievement: Hyper tuned Random Forest Model with AUC score of 0.75, Accuracy of 0.99 and Recall of 0.52.

• Domain: Stock Market | Tech Stack: MySQL | Nov 20

• Objective: To generate Buy/Sell/Hold signal for six stocks on a given date

• Solution: Implemented SQL Queries to define a user defined function which gave Buy/Sell/Holdsignalonagivendatevia10days’and20days’moving averages

• Key Achievement: Successful results & predicted behavior of stocks to help users invest