On TAURUS, a hybrid track-and-leg quadruped robot, I lead the perception and autonomy stack for autonomous rough-terrain traversal. I derived and verified 3-DOF-per-leg forward and inverse kinematics and built a MATLAB/Simulink simulation suite.
At Nitte Centre of Excellence for Applied AI, I built a rainfall-prediction pipeline for coastal Mangalore using nine years of daily weather data. I benchmarked five models, with the Extreme Learning Machine performing best at R² 0.91 and a 1.57% error.
I built an on-device home gym trainer that combines YOLO person detection, MediaPipe Pose, and depth-fused 3D landmarks on an OAK-D Lite. I also trained CNN and transfer-learned ResNet models to classify diabetic-retinopathy severity.
As an Automation Intern at Siemens Centre of Excellence, NITK Surathkal, I programmed PLC ladder logic for a traffic-light sequence and designed HMI screens. I work with Python, PyTorch, and TensorFlow, and my training includes robotics, machine learning, computer vision, and automation.
