KALUSHA AGUTI

Accomplished Machine Learning Engineer with expertise in Python and model evaluation, demonstrated at ChurchMMS. Proven ability to design and deploy robust fraud detection systems, enhancing accuracy through advanced techniques. Strong collaborator with a focus on delivering impactful data-driven solutions while ensuring transparency and reproducibility in all projects.

Fake News Detection System (Natural Language Processing)

• Built an NLP-based text classification system to detect misinformation using TF-IDF vectorization and machine learning classifiers
• Applied preprocessing techniques including tokenization, stop-word removal, and text normalization for improved accuracy
• Compared multiple models and optimized results through hyperparameter tuning and validation

Movie Recommendation System (Collaborative Filtering)

• Designed a recommendation engine using collaborative filtering to generate personalized movie suggestions
• Implemented similarity-based ranking methods to enhance recommendation relevance
• Evaluated system performance through user preference matching and recommendation accuracy

Chicago Traffic Accident Severity Prediction (Applied Machine Learning)

• Developed predictive models to classify accident severity using traffic and environmental datasets
• Conducted exploratory data analysis, feature selection, and model evaluation to improve predictive performance
• Produced visual insights and classification outputs to support transportation safety analysis

Customer Segmentation (Unsupervised Learning)

• Performed customer clustering using K-Means to identify distinct behavioral and spending groups
• Reduced dimensionality using PCA to improve interpretability of segmentation results
• Generated actionable insights for targeted marketing and customer analytics

Dog vs Cat Image Classification (Deep Learning)

• Built a convolutional neural network (CNN) model for binary image classification using TensorFlow/Keras
• Applied image preprocessing and augmentation techniques to improve model generalization
• Achieved strong classification performance on unseen validation data

House Price Prediction (Regression Modeling)

• Developed regression models to estimate housing prices based on structural and location features
• Performed data cleaning, transformation, and feature scaling to improve prediction reliability
• Assessed model accuracy using RMSE and cross-validation techniques

Rainfall Prediction (Time Series & Regression)

• Implemented predictive models to forecast rainfall patterns using historical climate data
• Applied regression approaches and performance evaluation metrics for environmental forecasting
• Generated outputs useful for climate monitoring and planning applications

Gold Price Prediction (Financial Forecasting)

• Built machine learning regression models to predict gold price trends using market indicator datasets
• Conducted feature engineering and model evaluation to improve forecasting accuracy
• Demonstrated application of ML techniques in financial analytics and prediction

• Humanitas Afrika - Full Scholarship for High School Education, 2012-2015
• SIUE Student Conference Travel Grant, 05/25