Back to projects
Machine LearningComputer VisionDefense
Defense Analytics Platform
Mission-critical defense analytics with computer vision and predictive modeling.
June 2022
About This Project
During my tenure at Quest Defense, I led the development of three mission-critical analytics projects for defense applications. The most significant was a predictive parts inventory system using Random Forest with isotonic regression calibration, reducing equipment downtime by optimizing spare parts availability. Another major initiative involved training a custom YOLO object detection model on synthetic imagery generated with Unreal Engine, enabling real-time analysis of drone reconnaissance footage. These projects operated under strict security requirements and earned me Employee of the Quarter recognition for exceptional performance in a high-stakes environment.
Tech Stack
Python
PyTorch
YOLO
Unreal Engine
scikit-learn
OpenCV
TensorFlow
Key Highlights
- Production-grade machine learning models
- Real-time computer vision processing
- Mission-critical reliability and security