• Built a cross-platform product authentication system for iOS and Android that identifies products from camera images and sends results with GPS metadata to backend services.

• Developed native mobile imaging and inference pipelines in Swift and Kotlin using Camera APIs, image processing, and on-device CNN inference.

• Delivered reusable mobile SDKs for iOS (xcframework) and Android (aar) to support partner integration and faster application development.

• Led data wrangling, dataset creation, model training, and TensorFlow Lite conversion for edge-deployed ML tasks including OCR, object detection, and semantic segmentation.

• Worked with OCR and vision stacks including PaddleOCR, EasyOCR, Tesseract, Keras-OCR, SSD, YOLO, and segmentation frameworks for embedded and mobile use cases.

• Integrated TensorFlow Lite with native C++ across Android and iOS to improve shared inference logic and cross-platform consistency.

• Diagnosed, maintained, and expanded a Linux embedded Qt/C++ product authentication application.

• Debugged and troubleshot production issues in OLED (Organic Light-Emitting Diode) inkjet printers using C# and the TwinCAT (The Windows Control and Automation Technology) suite, ensuring reliable operation and quality output

• Developed a time-based periodic maintenance system for critical printer components, including print head self-cleaning, nozzle validation, and mechanical positioning checks, to enhance longevity and performance

• Enhanced and maintained advanced digital image processing algorithms for blob detection in the Drop Measurement Unit (DMU), improving the precision of droplet size measurement

• Re-engineered the main software branch to integrate legacy production code with new features, ensuring backward compatibility and facilitating future enhancemen

Developed and maintained embedded systems and software for custom built equipment that transports, measures and processes glass substrates and Integrating tools to the Manufacturing Execution System

• Built QT 5.2.1 IDE (open source) on Odroid Mate 1.8 and developed Crane Authentication Reader UI for embedded application on Arm

• Cross-compiled custom kernel and touchscreen driver for Arm in Ubuntu 14.04

• Developed and maintained Windows-based embedded Crane Authentication Expert Mode Reader in C#

• Maintained and extended Crane Authentication User Mode Reader in WPF with Assisticant (Open Source Facade Design Pattern)

• Developed Android UI application in Java to drive embedded reader with RESTful API server on Windows

• Reverse-engineered and utilized commercial GPS device to detect Crane Authentication Reader coordinate