Rakib Ansari

Software Engineer

Deloitte•  June 2021 - Present

Involved in end to end data science pipeline. Performed date range experiment confirming the existing date range in production is optimal.  Executed VIF analysis : Identified moderate to strong multi-collinearity in top model features, leadingtoimproved feature selection for model training and improving model performance.  Updated the model and revised feature selection reducing the overfitting by 40%  Implemented a data split transformation from date-week to healthcare practitioner-led approach, significantly minimizing data leakage.  Implemented Pyspark code upgradation reducing run times as high as 60%.  Contributed in re-training the model which yielded in 7% total lift in market growth.  Oversaw model which helped in reducing over-fitting.  Improved model performance by 18% identifying filters and doing experiment on Hyper-parameter tuning.  Accomplished in predicting whether the doctors will prescribe the medicine or not to the enduser.  Delivered whole pipeline flow documentation.