Ajesh Saviour

Designed and implemented a fabric classification model using convolutional neural networks, optimizing it by transitioning to ResNet for improved accuracy of over 95%. Collaborated on the VirtuWear project, combining Stable Diffusion for fabric generation with virtual try-on modeling, earning an honorable mention award at DIGI+TCA Global Digital talent awards.

Designed and simulated robotic systems to enhance automation in various use cases, contributing to the lab’s ongoing projects in robotics and artificial intelligence. Collaborated on a team project that increased prototype functionality by 30%, leading to improved human-robot interaction models.

Designed and executed algorithms for constructing branched polymers and polymer gels, simulating their behavior under various conditions. Improved simulation efficiency by 20% through optimized LAMMPS scripting, achieving faster cohesive energy calculations for molecular systems. Analyzed simulation results to study properties such as mean square displacement (MSD) and cohesive energy, presenting findings with technical precision.