Ahteshaam Shaikh Email & Phone Number

Bengaluru, Karnataka, India

Developed a binary classification model of Logistic Regression, SVC and Random Forest Classifier to predict whether a customer will default on credit card payments or not.Performed EDA, Data Visualization and Feature Engineering on data for better modelperformance.Implemented SMOTE boosting technique to overcome the problem of class imbalance, carried out hyperparameter tuning and achieved Recall of 84.66% and ROC- AUC score of0.82. Plotted feature importance charts to identify the key issues causing customers to default on payments.

Developed a book recommendation system for amazon customers using memory and model based collaborative filtering by utilizing the description of book consumed and user interests.Implemented SVD based matrix factorization to obtain user_item interactions. Employed cosine distance to measure User-Item and Item-Item similarities. Created profiles for top active users by leveraging interaction strength with the recommended items and achieved test recall@5 of 23.76% and recall@10 of 30.47%

Developed a binary classification model using Logistic Regression, Random Forest Classifier and Naive-Bayes Classifier to predict whether the policyholders from the past year will be intrigued with the Vehicle insurance provided by the company.Designed the data pipeline to work on and performed EDA on behalf of the need for thecross-selling model building to come up with important insights. Performed feature encoding, implemented SMOTE boosting to oversample the minority class observations, carried out hyperparameter tuning and achieved AUC score of 0.95. Plotted feature importance charts to identify the important factors responsible for customer involvement.

Developed Linear Regression, XGBoost and Random Forest Regressors to predict the number of views of TED talk videos and reduced MAE to 0.2 million where target mean was around 2.6 million. Carried out feature engineering themed across speaker popularity, ted talk content, release day, popular topics, talk language and came up with the best performing features. Performed Hyperparameter tuning using GridsearchCV and RandomsearchCV to regularize thebias-variance trade-off.Identified the key factors driving views such as speaker profile, event category, Friday release; TEDx can consider these features to increase the views of their videos.

Navi Mumbai, Maharashtra, India

Mumbai, Maharashtra, India