● Developed ML models (XGBoost, CNN, Naïve Bayes, AdaBoost) for health condition classification with >90% accuracy.
● Automated disease diagnosis using OCR and PDF/image parsing, replacing manual doctor analysis.
● Built FastAPI backend to deliver AI-powered health summaries (GPT-4o), deployed with secure REST endpoints for mobile and kiosk use.
● Created real-time analytics dashboards (Matplotlib, Plotly) for demographic and health insights.
● Participated in Agile scrums and collaborated on full-stack integration.
Projects:
Staff Health Analytics Platform (github.com/sofaquitegud/ehealth).
● Built FastAPI-powered tool to analyze staff health metrics from CSV/PostgreSQL sources.
● Integrated OpenAI API to generate GPT-4 summaries and personalized health recommendations.
● Enabled dual-mode support (mobile/kiosk) with visualization, reporting, and demographic insights.
Disease Prediction from PDF/Image (github.com/sofaquitegud/syafiq-project).
● Developed Streamlit app combining EasyOCR and XGBoost to predict diseases from scanned medical reports.
● Delivered real-time predictions and feedback via intuitive UI.
Deep Learning for Medical Imaging (Best Project Overall @ InnovateX Tech Expo 2024).
● Enhanced CBCT dental scans with StyleGAN2-ADA to improve AI diagnostic performance.
● Trained ResNet-50 and VGG-16 for image classification; improved model F1-score significantly.
Travel Insurance Conversion Prediction (github.com/sofaquitegud/travel-insurance-analysis).
● Built pipeline using Random Forest to predict insurance conversions.
● Cleaned and merged JSON/CSV datasets; engineered features to handle class imbalance.
