• Developed a Python framework for automated data generation, and Recursive Neural Network’s (RNN) training and predictions for estimation of well’s productive layer’s oil flow rate using Distributed Temperature Sensing (DTS) data, as an internship project at Interpretive Software Products (ISP).
• Collaborated in a research project whose goal was to pinpoint the optimal multi-phase flow parameters characterizing the immiscible flow displacement utilizing a meta-heuristic Genetic Algorithm (GA) and an ensemble based iterative Kalman Filter (KF), resulting in accurate estimations of oil recovery with average relative errors as low as 6%.
• As a group member of an optimization project, worked toward the estimation of the underground reservoirs’ petrophysical properties through implementation of numerous meta-heuristic optimization algorithms in nonlinear regression analyses.
• As part of my master thesis at Sahand University of Technology, successfully trained, debugged, and tested feed-forward neural networks for accurate estimation of reservoirs’ petrophysical properties.
• Created a collective SQL database, comprising the seismic and salt well disposal data, and identified the most influential attributes on induced seismic activities in mid-continent United States using a Genetic-optimized Kmeans algorithm.