Data Scientist

Projects

Jul 2024 - Aug 2024

Developed a Convolutional Neural Network (CNN) model achieving 95% accuracy in classifying tomato plant diseases.
Engineered a web application for real-time image classification, processing over 1,500 user-uploaded images.
Improved disease diagnosis speed by 50%, significantly aiding agricultural productivity.

Jul 2024 - Aug 2024

Accomplished over 99% prediction accuracy in identifying the impact of review sentiment on rating scores and user engagement by developing a machine learning regression model trained on categorized review text and thumbs-up counts.
Boosted model performance as measured by R² and MAE by transforming user-generated content into sentiment scores (1–5) and performing extensive feature engineering and text preprocessing.
Delivered key insights into user behavior by classifying review content into positive, neutral, and negative, and visualizing trends using Python libraries like Matplotlib and Seaborn.

Jun 2024 - July 2024

Accomplished generation of 10+ critical business insights as measured by their alignment with ad hoc executive requests, by writing and optimizing complex SQL queries to support decision-making for a leading hardware company.
Improved data accessibility and clarity by structuring queries for sales trends, inventory analysis, and customer segmentation, simulating real-world challenges posed by a data analytics director.
Demonstrated both technical proficiency and business communication skills through the creation of clear documentation and rationale behind each SQL query.

Jun 2024 - Aug 2024

Accomplished 92% classification accuracy in identifying 10+ celebrities from .png images by training a Convolutional Neural Network (CNN) on a dataset of over 5,000 labeled images.
Improved model robustness and reduced overfitting by implementing data augmentation techniques and optimizing with early stopping and dropout layers.
Delivered a fully functional, user-friendly Streamlit web application that allows users to upload images and receive real-time predictions, enabling interactive deployment of deep learning models.

Jun 2024 - Aug 2024

Accomplished 95%+ diagnostic accuracy in predicting diabetes as measured by evaluation on test data by training and deploying a machine learning model using user inputs such as age, BMI, and glucose levels.
Enabled self-assessment for over 100+ users by designing an intuitive Streamlit web interface for real-time, accessible predictions.
Improved model reliability and reduced uncertainty by selecting key features and fine-tuning classification algorithms, leading to more confident health risk assessments.

July 2024 - Sep 2024

Implemented an NLP pipeline using advanced machine-learning techniques to classify news articles with over 98% accuracy.
Reduced false positives by 35%, ensuring reliable detection of misinformation.
Designed a web interface allowing users to analyze news content interactively, boosting engagement by 30%.

Sep 2025 – Dec 2025

Studied UNICEF and World Bank data to evaluate how malaria prevention efforts (ITNs) relate to child mortality across diverse regions and income groups.
Created an interactive global dashboard with geospatial maps and income-stratified visualizations to communicate health disparities.
Used Random Forest modeling and fairness metrics (Brier Score) to assess predictive reliability across income subgroups.
Identified a minimal direct relationship between ITN coverage and mortality (r = 0.13), highlighting socioeconomic inequalities as key drivers.
Provided evidence-based insights on equitable ITN distribution and health policy targeting.