Scam Call Detection Pipeline
Real-time ML pipeline detecting phone scams with 90%+ accuracy. Analyzes call patterns using K-Nearest Neighbors to process thousands of calls per second. Interactive dashboard delivers actionable insights including scam trends, confidence distribution, and confusion matrix.
Tech Stack:
- Streaming: Apache Kafka (KRaft), Apache Spark
- Storage: AWS S3, Snowflake Data Warehouse
- ML/Viz: Python (scikit-learn, PySpark), Streamlit
- Infrastructure: Docker