Azza Farichi Tjahjono

AI-based CPR training and monitoring system

• Built a realtime CPR quality evaluation system using sensor data and machine learning.
• Designed an ESP32 and ultrasonic sensor pipeline for compression-depth capture.
• Implemented FastAPI ingestion, realtime inference, ONNX export, and a monitoring dashboard.

AI-powered parking assistant system

• Combined parking slot monitoring with an LLM assistant for structured user interaction.
• Used function calling to fetch realtime parking availability from backend data.
• Built API endpoints for querying and parking information retrieval.

3rd place winner in an international data science competition

• Developed a deep learning segmentation model to identify flood areas from imagery.
• Enhanced a U-Net architecture with an attention module for stronger feature representation.
• Evaluated segmentation quality with Dice coefficient and IoU.