Current work

• SimBoard contributor, E3SM
  A web catalog of E3SM simulation runs: case metadata, provenance, and links to each run's diagnostics. FastAPI, PostgreSQL, and React, deployed at NERSC. app · docs · code
• UXarray lead developer
  Python library for analysis on unstructured climate grids, used at DOE labs and NCAR. I wrote its MCP server, which lets AI agents run UXarray analyses on HPC systems through Globus Compute.
  Details

  • Made a 300-million-cell, 63 GB mesh loadable in 306 seconds by deduplicating coordinates before building the grid (write-up).
  • Replaced a cross product in the spherical geometry kernels with error-free arithmetic after it drifted 0.68 m on a real Earth mesh (PR 1513).
  • 1.67× faster in a Numba hot path by removing hidden heap allocation (PR 1577, write-up).
  docs · code · MCP write-up
• E3SM AI platform contributor
  An assistant for E3SM documentation that cites its sources or declines to answer. I added Argonne's Argo gateway as a model provider.
• ncvis maintainer, DOE SEATS
  A NetCDF viewer for structured and unstructured grids, written in C++.

Earlier

• 2025–26 SFNO on Aurora. Ported a PyTorch climate emulator to Aurora's Intel GPUs and set up distributed training.
• 2017–25 CANDLE and IMPROVE. Hyperparameter search for cancer drug-response models, 10,000+ training runs on Summit, Theta, and Cori. R&D 100, 2023.
• 2019–24 FLASH-X. Asynchronous checkpointing with compression, cutting checkpoint overhead 40–70% on Summit. R&D 100, 2022.
• 2016–19 Urban ECP. A weather-to-building-to-CFD workflow for downtown Chicago.
• 2009–16 MeshKit. Reactor core mesh generation, as PI. A 101-million-element core mesh in about 7 minutes on 712 cores.

Tools

Languages

Python, C++, Fortran, Bash

Scientific Python

NumPy, Xarray, Numba, Dask, Polars, PyTorch

AI systems

MCP servers, agent evaluation, distributed training, hyperparameter search

HPC

MPI, HDF5, NetCDF, SZ3 and ZFP compression, Intel and NVIDIA GPUs, PBS, Slurm

Engineering

CMake, CI/CD, packaging (conda-forge, PyPI), Globus Compute, code review

Selected papers

• Jain, R., Jacob, R. Beyond Tool Execution: Evaluating Scientific MCP Interfaces with UXarray. AGENT4SC at IEEE eScience, 2026.
• Partin, A., …, Jain, R., et al. Benchmarking community drug response prediction models. Briefings in Bioinformatics, 2026.
• Jain, R., Tang, H., Dhruv, A., Byna, S. Enabling Data Reduction for FLASH-X Simulations. DRBSD-10 at SC24, 2024.
• Jain, R., Wozniak, J.M., Partin, A., et al. Cross-HPO: Optimizing Neural Networks for Cancer Drug Response. CAFCW at SC24, 2024.
• Wozniak, J.M., …, Jain, R., et al. CANDLE/Supervisor: A workflow framework for machine learning applied to cancer research. BMC Bioinformatics, 2018.
• Tautges, T.J., Jain, R. Creating Geometry and Mesh Models for Nuclear Reactor Core Geometries. Engineering with Computers, 2012.

Everything else is on Google Scholar.

Recent talks

• AMS 2027 Conservative zonal averaging on unstructured Earth-system grids
• SC26 Poster: Refuse to Guess, a declared precondition in the UXarray MCP server
• eScience 26 Evaluating scientific MCP interfaces with UXarray
• SciFM26 A UXarray MCP server for unstructured Earth-system meshes
• SC24 Tutorial: UXarray for unstructured climate data
• SciPy 2023 UXarray for unstructured climate data

Recognition

• 2023 R&D 100 Award, CANDLE
• 2022 R&D 100 Award, FLASH-X
• 2015 ATPESC scholar
• 2010 Best paper, International Meshing Roundtable
• 2004 First prize, low-budget car design contest, IIT Kharagpur

Service

• 2026 Program committee, AgenticAI4HPC at SC26 and AGENT4SC at eScience
• Ongoing DOE SBIR/STTR proposal reviewer

Background

• 2009– Argonne National Laboratory, Principal Specialist in Research Software Engineering
• 2023– University of Chicago, CASE, staff at-large
• 2020 M.S. Computer Science, University of Chicago
• 2009 M.S. Structural Engineering, Arizona State University
• 2006 B.Tech. Mechanical Engineering, IIT (ISM) Dhanbad