an engineer:
with a knack for solving problems and following curiosity.

Computer-vision research for road intelligence and autonomous-driving use cases.

• Built road-sign detection pipelines and evaluated AutoML approaches on Google Cloud.
• Benchmarked transfer-learning architectures including ResNet50, Xception, and InceptionResNetV2.

Risk intelligence for a university-wide vendor ecosystem.

• Redesigned vendor-risk prediction for 28,000 vendors, reaching a 0.91 F1-score.
• Reduced false positives by 15% through XGBoost deployment and A/B validation.
• Automated alerts and Power BI reporting, saving more than 20 hours each week.

Efficient deep learning for chest X-ray diagnosis.

• Developed a low-FLOP approach for COVID-19 and pneumonia detection.
• Used U-Net encoder-decoder architectures to double training speed.
• Deployed Micronet M3 with 99.3% accuracy and 99.31% F1-score.

Indoor positioning using wireless fingerprints and inertial signals.

• Built k-NN and Random Forest positioning models with two-to-three meter precision.
• Designed a real-time RSSI localization algorithm reaching 0.98 accuracy.