Projects

Speech Emotion Recognition

COVID-19 Chest X-ray Image Classification

Built a Python-based ML model to classify emotions from speech using the RAVDESS dataset. Extracted MFCC, chroma, and mel-spectrogram features with librosa, showcasing audio processing and feature engineering skills. Implemented an MLP classifier, leveraging neural network design and tuning expertise, achieving high accuracy. Demonstrated analytical and problem-solving abilities, with applications in mental health and customer service."

Created a Python ML model to predict COVID-19 from chest X-rays using a managed dataset of image IDs and labels. Processed images with Excel macros, showcasing data management skills, and resized them to 200px with image processing tools. Converted 3D images to 2D arrays for analysis, demonstrating preprocessing expertise. Built and evaluated a model, achieving accurate classification, with applications in early healthcare detection."

Developed an ML model to predict SBA loan status (charged off or not), tackling missing data and mixed types (numerical, categorical, dates). Preprocessed data, addressing imbalance with feature engineering for meaningful inputs, showcasing data handling skills. Built a classification model, evaluated with accuracy and metrics, demonstrating expertise in preprocessing, modeling, and performance analysis."

Loan Status Prediction

Telecom Company Churn Analysis

Built an ML model to predict telecom customer churn using contract, service, and demographic data. Encoded categorical features (e.g., contracts, payments) with label and one-hot encoding, addressing imbalance. Compared Logistic Regression, Random Forest, and XGBoost models, evaluated via accuracy, precision, recall, and ROC-AUC. Applied scaling and feature selection, fine-tuning for retention insights, showcasing preprocessing and modeling skills.