At Aligner AI, I author reinforcement-learning tasks in seismic retrofit and structural stability, including OpenSees damper and steel-brace retrofits and a pony-truss chord-buckling retrofit. Each task includes a verified reference solution and automated grader, and I check model answers against ACI, AISC, Eurocode, and ASTM provisions.
My research examines earthquake damage and the durability of cement-stabilized rammed earth. I built a hierarchical Bayesian model to infer shaking intensity and building vulnerability jointly from district damage counts after the 2005 Kashmir earthquake, and derived fragility functions for Himalayan masonry.
At Sharif University of Technology, I extended rammed-earth testing to 365 days and analyzed published durability results across studies and laboratories. I also completed an experimental thesis using Himalayan soil and designed and built a structural spaghetti bridge that held more than 140 kg.
