What I Work On

A PPO agent (with SAC as a comparison baseline) trained on a custom Gymnasium environment that wraps an Aspen Plus simulation of the Haber–Bosch process. The agent learns to set operating variables — temperature, pressure, recycle ratio, feed composition — to minimize energy intensity per kilogram of ammonia produced, subject to converter and downstream constraints.

The motivation: green ammonia is a strong candidate carrier for hydrogen, but the energy cost of synthesis remains the bottleneck. Classical optimization handles steady-state design well, but struggles with dynamic operating conditions tied to intermittent renewable input. Reinforcement learning offers a path through that.

Investigating how applied magnetic fields can enhance the kinetics of electrocatalytic water splitting. The hypothesis is that spin-polarized electron transfer at ferromagnetic catalyst surfaces can lower the overpotential required to drive the oxygen evolution reaction — the rate-limiting half of the electrolyzer.

Working on synthesis, characterization, and electrochemical testing of candidate catalysts under varying field strengths.

Synthesizing graphene-supported titanium oxide nanocomposites and characterizing them as electrocatalysts for the hydrogen evolution reaction. Characterization stack: SEM for morphology, XRD for crystal structure, cyclic voltammetry for electrochemical behavior, TGA for thermal stability, and IR spectroscopy for functional groups.

Ibdaa 2024 finalist · IAEA Special Award · Aramco Special Award

A density functional theory study comparing Al, Si, and B doping of graphene sheets for selective HCN gas sensing. Computed binding energies, charge transfer, and electronic structure changes to evaluate each dopant's sensitivity and selectivity.