Cosmology · Astrophysics

Utkarsh
Giri

I'm a postdoctoral fellow at Caltech, working on cosmological surveys and statistical methods for extracting signals from large-scale astrophysical datasets. On the side, I like tinkering with machine learning.

Publications GitHub ↗ Scholar ↗
Utkarsh Giri
Experience
Jun 2024 – Present
California Institute of Technology
Postdoctoral Fellow
  • Working on data analysis and inference pipelines for the SPHEREx satellite mission.
  • Developing methods to extract cosmological signals from noisy, high-dimensional survey data.
Sep 2021 – Jun 2024
University of Wisconsin-Madison
Research Associate
  • Developed a neural network-enhanced estimator for primordial non-Gaussianity.
  • Built simulation-to-inference pipelines using N-body simulations and halo-model analytics.
  • Contributed to Bayesian analysis of fast radio burst signals as part of the CHIME/FRB Collaboration.
Education
2016 – 2021
Perimeter Institute & University of Waterloo
Ph.D. in Physics
  • Thesis: Reconstructing Cosmic Velocities with the Kinetic Sunyaev-Zeldovich Effect
  • Advisor: Kendrick M. Smith
2010 – 2015
Indian Institute of Science Education and Research, Pune
BS-MS Dual Degree
  • Thesis: Lepton Flavour Violation in Supersymmetry
  • Advisors: Rohini Godbole and Sudhir Vempati
Open source code

fastBK

MPI-parallel bispectrum computation. Independently validated by two subsequent groups.

jaims

JAX-based affine-invariant MCMC sampler for Bayesian inference.

zinc

Dockerized Zel'dovich & 2LPT initial condition generator for N-body simulations.

remotenb & tunnelboard

Remote tunneling tools for Jupyter notebooks and TensorBoard.

Side projects
Projects
All projects →