Nilson Chapagain
ABOUT ME
My training is in statistics, with interests in offline policy learning, causal inference, and statistical learning for sequential decision-making.
But a field of study is only one way to introduce a person. Much of what shapes me happens Shows an explanation. Press Arrow Down to enter the preview, or Escape to dismiss it.: a question that follows me home, a conversation I reconsider, a small decision nobody sees.
Above an early sketch of his Shows an explanation. Press Arrow Down to enter the preview, or Escape to dismiss it., Darwin wrote, “I think.” I like the room those words leave for doubt. Shows an explanation. Press Arrow Down to enter the preview, or Escape to dismiss it., too, recognized the value of remaining with uncertainty rather than rushing toward an answer. That is a patience I hope to cultivate: holding an idea closely enough to examine it, and loosely enough to change my mind.
Yet learning is never entirely solitary. Marcus Aurelius began his Shows an explanation. Press Arrow Down to enter the preview, or Escape to dismiss it. by remembering what he had learned from others. I like that an account of the self could begin with gratitude. Whoever I become will carry traces of someone’s patience, someone’s difficult question, someone who made room for me before I knew how to ask. I hope to recognize those gifts and pass them on.
EDUCATION
Texas A&M University
Ph.D. in Statistics
Stony Brook University
B.S. in Applied Mathematics and Computer Science
Publications & Preprints
Cao, T. H., Chapagain, N., Lee, H., Thi, P. N., & Thieu, N. N.[PDF]
Optimal Control of Several Motion Models
Journal of Optimization Theory and Applications•2025
Poudel, B., Chapagain, N., Kumar, A., & Jiang, X.[PDF]
Reasoning-Search-Augmented LLMs: An Operational Taxonomy, Benchmark Analysis and Controlled Evidence Synthesis
Laha, N., Chapagain, N., Cicherski, V., & Sonabend-W, A.[PDF]
Dynamic Treatment Regimes via Smooth Surrogate Loss for Arbitrary Number of Treatments per Stage
Chapagain, N.[PDF]
Support-Aware Offline Policy Selection for Two-Stage Recommender Systems
Software
Chapagain, N., & Pati, D.[CRAN][PyPI][Documentation]
Variable Selection using Shrinkage Priors (VsusP)
Chapagain, N.[PyPI][Documentation]
QuantileMix: Quantile g-computation in Python
Chapagain, N.[PyPI][Documentation]
KernelMix: Bayesian kernel machine regression in Python
Chapagain, N.[PyPI][Documentation]
AShrink: Adaptive shrinkage in Python
Chapagain, N.[PyPI][Documentation]
JointShrink: Multivariate adaptive shrinkage in Python
Under Review
Chapagain, N., Li, K., & Selukar, S.
A bias-variance tradeoff in time-dependent Cox models with duration-dependent effects
Chapagain, N., Wong, R., & Laha, N.
Doubly Robust Dynamic Treatment Regimes via Simultaneous Direct Search
Work in Progress
Chapagain, N., Bhattacharya, S., & Laha, N.
Multi-Stage Dynamic Treatment Regimes under Network Interference
Cao, T. H., Chapagain, N., & Jeong, S. M.
Calibration of Neural Sweeping Dynamics under Hard Constraints: Strong Discrete Approximation and Contact-Aware Adjoint Sensitivity

