●Developing LLM-driven, role-specific AI characters (patients, doctors, staff, family) for intelligent ICU digital twins.●Developing contrastive and counterfactual explainable AI methods to interpret results from Transformer models for time series data in healthcare.●Establishing real-time data integration pipelines for non-protected health information (PHI) streaming between the ICU and its digital twin.

●Developed a dynamically weighted logistic function based multiobjective optimization framework for redirecting gestures with pose preservation between remote environments in mixed reality. The framework had an interaction error reduction of 34% compared to the current state-of-the-art.●Developed a mesh deformation based method to map real time locomotion between remote environments of any irregularly shaped topology in mixed reality.●Developed a Lowner ellipse based obstacle avoidance method for smooth and naturalistic traversal of avatar motion around any irregularly shaped obstacle. ●Developed an augmented reality application where users can remotely interact with other users and objects in their mutual environment through their photorealistic avatars. The application was developed using Unity and Photon networking engine and was tested on 60 participants with significant results. ●Classified sleep stages in children with Rett syndrome (RTT) using physiological and accelerometer signals obtained from a smartwatch sensor. The SVM-based machine learning model can differentiate between awake, non-REM sleep, and REM sleep in RTT children with an accuracy of 85.1%.●Developed a Random Forrest stress detection classifier for older adults with 75% detection accuracy. The classifier model can be used to dynamically adjust the difficulty levels in closed loop HCI intervention systems for older adults. ●Helped write a National Science Foundation (NSF) Smart and Connected Communities (S&CC) grant # 2225890 “Reducing Loneliness for Long Term Care Older Adults through Collaborative Augmented Reality” and National Institute of Health (NIH) R21 grant # 1R21AG078480-01 “Reducing Loneliness of Older Adults in Long Term Care Facilities through Collaborative Augmented Reality” based on my thesis with my advisor Dr. Nilanjan Sarkar as PI and my dissertation committee member Dr. Lorraine Mion as Co-PI that was funded for $2.25 million and $433,495 respectively.

●TA for the ECE Capstone projects for the Fall 2017 and Spring 2018 semesters. Responsibilities included evaluating and grading Capstone proposals. ●Involved in helping students troubleshoot with projects using Arduino, Raspberry Pi and evaluating their simulations on LABVIEW.

●Developed an accelerometer sensor system for recliner chairs that can noninvasively extract heart and respiration rates from the occupants' ballistocardiogram signals. ●Tested the sensor system with 45 subjects, having an average age of 78.8 years at Tiger Place, Columbia, Missouri. The mean heartbeat error for 45 subjects was 0.6 msec with an average error rate of 3.6%. Similarly, the mean respiratory breath error was 4.2 msec with an average detection error rate of 6.25%.

●Developed a deep neural network to classify seizure vs seizure-free EEG signals based on the features obtained from Variational Mode Decomposition (VMD) of the signals . An accuracy of 98.4% was obtained in the classification. ●Invited to write a book chapter for the developed automated VMD based seizure classifier in Deep learning, machine learning and IoT in biomedical and health informatics, CRC Press

●Developed a code using NS-2(Network Simulator 2) that compares AODV (Ad-hoc On Demand Vector and DSDV (Destination Sequenced Vector Routing) with different node configurations.●Simulated TCP and UDP protocols using NS-2 with emphasis on their bit rates.

●Analyzed and helped in the maintenance and working of the Distributed Control System (DCS) and Programmable Logic Controller (PLC) systems used in the fail-safe mechanisms of the refinery.●Conducted in-plant studies of wireless transponders, pyrometers, pressure gauges and gas detectors used during running, maintenance and repair operation of the refinery.