Giriraj Pawar

Troubleshot bugs in C++ to ensure software performance, resolving system defects effectively. Conducted thorough code reviews, collaborating with internal/external teams to maintain code quality. Developed KPIs to compare feature branches with develop during regression testing, ensuring stability. Achieved more than 90% code and 80% branch coverage through rigorous unit testing, ensuring robust functionality. Collaborated in agile methodologies for efficient software application lifecycle management.

Conducted a literature review on image-to-image translation and domain adaptation to reduce gaps between synthetic and real images. Prepared synthetic and real datasets, applying normalization, scaling, and augmentation for better robustness. Built a pipeline using CycleGAN to convert synthetic images to realistic counterparts. Trained and optimized CycleGAN with cycle consistency loss to reduce domain gaps. Evaluated image quality with a classifier, achieving 65% accuracy in reducing the domain gap.

Implemented a cross-platform C++ application for the Vector PCAN-USB Pro device to enable real-time processing of CAN and LIN data via a USB interface. Formulated unit tests to achieve over 95% code coverage and 90% branch coverage in the CI/CD pipeline, ensuring robust validation and reliability of real-time CAN and LIN data processing.

Developed a multi-threaded cross-platform C/C++ middleware application to analyze OEM thermal designs using a GUI/WebUI. Enhanced command-line functionality for seamless integration with other tools, introduced new features, and played a key role in documentation, debugging, and bug resolution.

Maintained and optimized C/C++ applications for Text-to-Speech (TTS) conversion, integrating the iFLYTEK library to produce natural-sounding speech on embedded devices. Implemented PCM sample playback and recording functionality using the ALSA library and resolved critical issues, addressing crashes and improving system stability.

Computer Vision Real Time Systems Image processing and pattern recognition Deep learning Machine learning