- Large Language Models: Utilizing Large Language models for automated vulnerability analysis and developing safe and reliable systems.- Physics-Informed Machine Learning: Utilizing principles from physics to enhance the effectiveness and interpretability of machine learning models.- Informed Decision-Making: Leveraging data-driven insights and analytical techniques to make informed and strategic decisions.- Human-Machine Symbiosis: Designing systems where humans and machines collaborate seamlessly, optimizing each other's strengths for enhanced performance.- Trustworthy Systems for Human-Machine Interaction: Developing systems that prioritize safety, reliability, and user trust in interactions between humans and machines.- Modeling and Simulation of Physical Phenomena: Applying computational models to simulate and analyze real-world physical processes, aiding in understanding and prediction.- Application of Assume-Guarantee Contract-Based System Design: Implementing design methodologies that ensure system components meet specified requirements, enhancing system reliability and performance.

• Developed predictive tools for informed decision making• Developed ML based models and runtime efficient multi-thread algorithms for real-time decision• Created an ML based framework for Radar Doppler Image sequencing • Designed a framework for global sensitivity analysis for high fidelity simulation runs• Developed a workflow for analyzing historical drifts in simulation results

• Developed predictive mathematical tools for system identification, reserve estimation, and multiphysics fluid monitoring for de-risking exploration and field development planning.• Designed and developed a new probabilistic method for radius of investigation for use in deep pressure transient test design. The application was implemented in a commercial software for real-time data analysis to facilitate communication and optimize efficiency for effective operations. • Created a machine learning based proxy model for time series prediction and analysis of pressure measurements using GRU-RNN with PyTorch - CUDA supported by Bayesian optimization. New model reduced the CPU time in pressure prediction by 2-orders of magnitude.• Established a cost-efficient workflow for multi-physics monitoring designing tool for informed "go" or "no-go" decision making under uncertain formation properties on the system and the sensors. Implemented solution in a commercial software platform (PETREL) used by thousands today.

• Developed a general-purpose “Quick-look” solution for modeling the thermodynamic properties of hydrocarbons – brine mixture and quick assessment of multiphysics fluid monitoring. • Conducted research on developing multiphysics fluid monitoring methods for carbon dioxide storage and enhanced oil recovery.• Served as technical lead from Schlumberger with Lawrence-Berkley Lab on the feasibility of carbon dioxide injection into faults for enhanced characterization of faults in geothermal systems.• Investigated the effect of CO2 dissolution on bulk modulus and sound speed of hydrocarbons and quantified the effect of dissolution on seismic monitoring of carbon dioxide in oil recovery.

• Built accurate mathematical models for calculating the acoustic properties of CO2 suitable for oilfield and geological CO2 storage applications.• Developed an NPV based algorithm for real-time economic assessment of CO2 storage in saline aquifers and cost-effective well-placement and scheduling.• Implemented a new mathematical model for two-phase three-component flow in porous media for use in CO2 storage in saline aquifers.• Studied the permanent geological CO2 storage in subsea below the negative buoyancy zone (in collaboration with Harvard University).• Explored the effect of capillary pressure hysteresis on CO2 flow in saline aquifers and identified capillary pressure hysteresis as another trapping mechanism retarding the movement of CO2 in saline aquifer.• Technical contributor in a cross-industrial project on reservoir characterization through joint inversion of formation parameters from permanent electrode array measurements.• Developed fluid substitution model for use in seismic monitoring in CO2-EOR and CO2 storage and multiphysics inversion of formation properties.• Researched multiphysics fluid monitoring methods for improving oil recovery in CO2-EOR and verifying the permanence of geological CO2 storage.

Introduced a new mathematical method for detecting buried objects and their shapes: an application of Phase-field models to industrial problems

Over the course of my graduate years at the University of Pittsburgh, I taught undergraduate level calculus and ordinary differential equations courses.