Machine Learning
Data Analysis & Model Development
AI & Machine Learning
11 min read
March 2026
Introduction
Machine learning enables computers to learn patterns from data and make intelligent predictions. I've built end-to-end ML pipelines including data preprocessing, feature engineering, model training, evaluation, and deployment. My work spans classification, regression, and deep learning tasks.
Key Learnings
Data Preprocessing
Cleaning, normalizing, and transforming raw data with handling of missing values, outliers, and feature scaling.
Feature Engineering
Creating meaningful features from raw data using domain knowledge and statistical techniques.
Model Training & Tuning
Training machine learning models with hyperparameter optimization and cross-validation for robust performance.
Model Evaluation
Comprehensive evaluation using appropriate metrics like precision, recall, F1-score, and confusion matrices.
Tools & Technologies
Python
Primary language for ML development with extensive library ecosystem.
Pandas
Data manipulation library for loading, cleaning, and transforming datasets.
NumPy
Numerical computing library for array operations and mathematical functions.
Scikit-Learn
Machine learning library with algorithms for classification, regression, and clustering.
PyTorch & TensorFlow
Deep learning frameworks for building and training neural networks.
How I Used This in Projects
→DementiaInsight - Non-Medical Dementia Risk Classifier (Dec 2025)
Developed an automated ML pipeline to predict dementia risk using non-medical features. Implemented data processing, model training and evaluation, and a CLI prediction tool. Achieved strong performance with LightGBM and was a finalist at the ModelX Inter-University Hackathon.
→MedPredict - Medical Cost Prediction Model (Nov 2025)
Developed a Random Forest regressor to predict medical insurance costs using lifestyle and demographic indicators. Deployed via Streamlit to provide a lightweight interactive web interface.
→Image Classification with Deep Learning
Implemented CNN using PyTorch for image classification. Trained on custom dataset with data augmentation and achieved 92% validation accuracy.
Skills & Tags
Want to explore more?
← Back to Portfolio