Developing and monitoring neural networks for defect detection.Improving model performance with pruning and explanability.

Performed state-estimation and observability analysis for non-linear systems.

Educated students on key programming concepts including data structures, algorithms, and object-oriented programming principles.

● Labeled datasets, trained and evaluated a YOLO v5 model for detection of endangered species to prevent poaching.● Performed data augmentation and hyperparameter tuning to improve the mAP to 70%.● Implemented ML algorithms to improve PID control for vision-based motion models, reducing tracking errors by 80%.

● Designed a PID control algorithm for differential steering and prevention of slip of a rover on sandy terrain.● Developed SLAM algorithms for coastline marking using satellite images for autonomous navigation.● Tuned the navigation parameters and performed 3D point cloud filtering for obstacle avoidance using ROS.● Documented the performance in real and simulated environments and reduced the computation time by 10%.● Created an application to monitor the node status and display live video stream from the rover using AWS Kinesis.

● Designed a self-actuated deployment mechanism for solar sails to reduce the area before deployment by 30%.● Developed a prototype for an artificial limb and to control its movement using state estimation

• Point cloud filtering and surface reconstruction of 3D ship maps.• Pressure drop analysis and forecasting.• Validated the control response time for fire suppression measures on offshore helidecks.

• Investigated the reliability of a gasoline direct injection (GDI) system for an all-weather automobile.• Gained hands-on experience of dismantling and re-assembling an IC engine.