GW-BSE Study of Chromium Sulfur Bromide
First-principles many-body perturbation theory study of electronic, optical, and excitonic properties in the layered magnetic semiconductor CrSBr using GW-BSE methods.
Research
Current
- One of my current main interests is computation in recurrent neural networks. This is includes (but is not limited to): sequence generation, non-normality, random matrix theory, noise in neural networks, low-rank networks, importance of time scales
- Dynamical mechanisms of compositional computation
- Interpretability of large language models, especially: symbolicity, reasoning and syntax
- Binding and compositionality in the brain, specifically in language processing
Past
Machine-Learned Interatomic Potentials for BaTiO₃
Developed machine-learned interatomic potentials (MLIPs) for barium titanate to reduce computational cost of ab-initio molecular dynamics.
Angle-Resolved Photoemission Spectroscopy & Ultrafast Spectroscopy
Performed ARPES measurements on 2D materials to map electronic band structure and explore momentum-space exciton dynamics. Designed and built a fourth-harmonic generation (FHG) setup for deep-UV pump-probe experiments, optimizing nonlinear frequency conversion in BBO crystals.
THz LC-Resonators for Ultra-Strong Light-Matter Coupling
Fabrication and characterization of terahertz metamaterial LC-resonators for ultra-strong light-matter coupling.
In Vitro Studies of Actin-Based Motility
Experimental and theoretical study of actin-based motility using VCA-coated polystyrene beads.