---
layout: '@/layouts/Doc.astro'
title: '🎉 Happy New Year!'
date: 2026-01-07
description: 'A New Year update summarizing ongoing projects including AuroraGPT, AERIS, and other involvements at Argonne.'
---

- [Ongoing Projects](#ongoing-projects)
- [Additional Involvements](#additional-involvements)

I’d like to try and post more this year.

Ideally these would be less-polished, more-frequent updates on what I’m
thinking about / working on.

## Ongoing Projects

- [**AuroraGPT**](https://auroragpt.anl.gov): Large Language Models for
  Scientific Applications on leadership-class supercomputers[^1].  
  Additional details can be found in some of my recent talks:
    - [AuroraGPT: Training Foundation Models on Supercomputers](https://samforeman.me/talks/2025/12/16/)
    - [Training Foundation Models on Supercomputers](https://samforeman.me/talks/2025/10/24/)
- [**AERIS**: Argonne Earth Systems Model for Reliable and Skillful
  Predictions](https://arxiv.org/abs/2509.13523) (Hatanpää et al. (2025)).  
  Foundation models for Earth system science, pushing on coupled
  modeling, uncertainty, and long-horizon prediction.  
  Additional details can be found in some of my recent talks:
    - [AERIS: Argonne’s Earth Systems Model](https://samforeman.me/talks/2026/10/08/)
    - [Training Foundation Models on Supercomputers](https://samforeman.me/talks/2025/10/24/)
- 🍋 [`ezpz`](https://saforem2.github.io/ezpz): A growing collection of
  utilities for launching, instrumenting, and debugging distributed jobs
  on real HPC systems.  
  This started as glue code and turned into infrastructure. A dedicated
  post is coming.
- [**Genesis Project**](https://genesis.energy.gov/)
    - [The American Science Cloud
      (AmSC)](https://american-science-cloud.github.io/amsc-site/): A
      Platform for Transformative Science

      > [!NOTE]+ AmSC
      > The American Science Cloud (AmSC) is a Platform for
      > Transformative Science that empowers scientists, engineers, and
      > researchers to harness the power of AI, data, and computing to
      > advance science, engineering, and energy missions.

    - The Transformational AI Models Consortium (ModCon): Cornerstone of
      the Genesis Mission’s AI models and data efforts.

      > [!INFO]+ ModCon
      > Cornerstone of the Genesis Mission’s AI models and data efforts,
      > will build and deploy self-improving AI models that advance science,
      > engineering, and energy missions by harnessing DOE’s unique data, facilities,
      > and expertise. Selected teams will develop foundational capabilities needed
      > across multiple scientific and engineering domains.

## Additional Involvements

- [**DeepSpeed Technical Steering
  Committee**](https://www.linkedin.com/feed/update/urn:li:activity:7412334024578375680/)[^2]

    <details is-="accordion">
    <summary>More info</summary>
    The DeepSpeed Technical Steering Committee (TSC) is responsible for
    guiding the overall direction and development of the DeepSpeed project.
    </details>

- [**CPSC**](https://imfisk.github.io/CPSC/): Member of the Coordinating Panel for Software and Computing.  
  <details is-="accordion">
    <summary>More info</summary>

    The Coordinating Panel for Software and Computing (CPSC) serves as a
    forum for the U.S. high energy physics (HEP) community to address
    shared challenges in scientific computing.
   
    Hosted by the Division of Particles and Fields (DPF) of the American
    Physical Society(APS), the panel brings together researchers,
    developers, institutions, and industry partners to strengthen the
    software and computing ecosystem that underpins modern HEP research.
    Through coordination, advocacy, and community-building, CPSC works to
    foster innovation, support career development, and ensure the
    computing infrastructure evolves to meet the demands of current and
    future experiments.
    </details>

<div id="refs" class="references csl-bib-body hanging-indent">

<div id="ref-stock2025aeris" class="csl-entry">

Hatanpää, Väinö, Eugene Ku, Jason Stock, et al. 2025. _AERIS: Argonne
Earth Systems Model for Reliable and Skillful Predictions_.
https://arxiv.org/abs/2509.13523.

</div>

</div>

[^1]: More on this soon!

[^2]:
    [Roadmap
    discussion](https://github.com/deepspeedai/DeepSpeed/issues/7705)