Hi, I'm Yubraj, a physics and mathematics undergraduate at Duke, originally from Nepal. Right now I'm chasing three questions: what a quantum simulation really has to keep when we only care about one observable, what a CT measurement can and can't tell us, and how few sensors it takes to work out where air pollution comes from. Around that I build small scientific tools, mentor high school students in Nepal through Unnchai, and, away from the screen, play chess and cook.
You can read more about me, browse my work and notes, skim my CV, find me on GitHub and Google Scholar, or just get in touch.
Research · Preprint
What needs to be simulated when we only care about a particular observable?
Readout-aware approximation in structured operator bases: symmetries, Pauli propagation, and weighted commutator graphs.
Research · Published
How can we vary organ size while keeping a generated abdomen anatomically coherent?
Sequential, volume-conditioned diffusion for 3D anatomy.
Research · Ongoing research
What can spectral CT measurements tell us about material composition—and what remains unidentifiable?
Differentiable spectral CT forward modeling, gradient-based inverse optimization, and identifiability analysis.
Research
How much can sparse measurements reveal about pollutant transport and sources?
PDE simulations, source inversion, and sensor-placement experiments.
Things I’ve learned, questions I’m following, and resources worth sharing.
My journey from Nepal to Duke, and some thoughts on college applications.
Resources and guidance for the International Olympiad on Astronomy and Astrophysics.
A small collection of machine-learning courses and resources.
Good questions belong to everyone.
Through Unnchai, I mentor high school students in Nepal on quantum algorithms. I care about making learning resources and educational opportunities easier to find—and leaving room to explore beyond formal research.
Community & education