Rohith Shinoj Kumar

Hi! I am a Computer Science and Engineering graduate from NITK, specializing in designing and deploying robust AI systems for complex, real-world applications. My work focuses on integrating advanced deep learning architectures and robust control theory through strong engineering and mathematical principles to ensure reliability in critical domains.

My research has centered on creating reliable AI tools for trustworthy AI and AI safety applications, ranging from developing an H-Infinity enhanced "control-aware" CNN-LSTM for arrhythmia detection from noisy heart audio to building a semi-supervised Pyramidal Vision Transformer (PVIT) for consistent Ejection Fraction quantification in echocardiography. This work is complemented by my experience from research assistantships at NITK and high-performance pipeline engineering at the Centre for Development of Telematics, India's national telecom R&D centre.

In my role as a Scientist at C-DOT, I design performance-critical kernel modules and optimize DL-based packet flow pipelines to work at throughputs of up to 700,000 packets/sec, which has provided me the opportunity to integrate cutting-edge research with highly optimized and quantized deployed systems. I am passionate about tackling complex problems where reliability, efficiency, and real-world deployment on constrained devices are most critical, with a particular interest in Healthcare AI applications.