Time-Series Anomaly Detection API
Time-series anomaly detection system built with rolling statistical features and an Isolation Forest model, exposed through a FastAPI service with careful warm-up handling and Dockerized deployment.
I’m a Data Scientist and AI Engineer with experience building end-to-end machine learning systems, automation workflows, and data-driven products across sectors including network automation, fintech, and proptech.
My work spans machine learning and deep learning, with hands-on experience in NLP, computer vision, time-series modeling, and survival analysis.
Time-series anomaly detection system built with rolling statistical features and an Isolation Forest model, exposed through a FastAPI service with careful warm-up handling and Dockerized deployment.
Developed multi-modal survival prediction models using clinical, genomic, and imaging data with CoxPH, DeepSurv, and DeepHit, plus SHAP-based explainability.
Built a CNN to classify emergency vs non-emergency vehicles, achieving 90% accuracy.
Time-series analysis and forecasting of UK road accident trends.
Demographic analysis and visualization of census-style data.
Regression models to predict video game sales performance.
NLP models for disaster tweet classification using BiLSTM & GRU.