I'm a Machine Learning Engineer with hands-on experience in fraud detection, computer vision, and predictive analytics.
I build end-to-end ML systems with a strong focus on real-world impact, especially imbalanced classification, probability-based evaluation, cost-aware decision making, and production-oriented model serving.
- Machine Learning: Classification, Imbalanced Learning, Threshold Optimization, Model Evaluation (PR-AUC)
- Models: XGBoost, Random Forest, Logistic Regression
- Deep Learning: CNNs (TensorFlow / Keras)
- Data Processing: Pandas, NumPy, Feature Engineering
- Computer Vision: OpenCV
- Tools: Python, FastAPI, Git, GitHub, Jupyter
- Built an end-to-end machine learning system for highly imbalanced fraud detection (≈0.17% fraud rate)
- Evaluated Logistic Regression, Random Forest, and XGBoost using PR-AUC and probability-based metrics
- Performed model-specific threshold tuning to balance fraud recall and false positive alerts
- Reduced false positive alerts by approximately 94% while maintaining ~86% fraud recall, significantly lowering expected operational cost
- Selected XGBoost as the final model using cost-based evaluation and expected financial loss analysis
- Served the trained model via a production-style FastAPI inference service with strict feature schema enforcement
- 🔗 ML Pipeline
- 🔗 Inference API
- Built a CNN-based CAPTCHA recognition system trained on a large-scale dataset (123K+ samples)
- Applied image preprocessing, augmentation, and inference optimization techniques
- Prepared trained models for inference and future deployment
- 🔗 View Project
- Developed a real-time face mask detection system using CNN and OpenCV
- Integrated temperature sensor readings with automated gate control logic
- Achieved 99% validation accuracy under controlled conditions
- 🔗 View Project
- Built a machine learning model to predict customer churn using structured telecom data
- Applied preprocessing, feature engineering, and model evaluation techniques
- Delivered actionable insights to support data-driven retention strategies
- 🔗 View Project
- GitHub: msaad-dot
- LinkedIn: Muhammed Saad
- Email: [m.saad7@outlook.com]