akougkas.io

Building GRC

A research center for HPC data systems at Illinois Tech.

The Gnosis Research Center, co-founded with Xian-He Sun in 2023 and now home to research faculty, engineers, and a dozen PhD students.

View the code
Anthony Kougkas with GRC students in front of the SC24 sign in Atlanta.
Atlanta, November 2024. The GRC team at SC24.

The question

What would a research environment look like if it treated people, software, and ideas as parts of the same system?

When I arrived in Chicago in 2013, high-performance computing at Illinois Tech was mostly one faculty laboratory, the Scalable Computing Software lab that Xian-He Sun had run since 1999. I did my PhD there, led its research and student development from 2019, and by 2023 the work had outgrown a lab. That year Xian-He and I co-founded the Gnosis Research Center.

GRC grew from a practical need. Ambitious systems work takes longer than a semester and involves more than a paper: repositories, releases, machines, proposals, documentation, mentoring, collaborations, and the confidence to try a direction that might fail. Those pieces compound only when someone tends the connections between them.

The organizational bet

My bet is that a center can be rigorous without becoming impersonal, and entrepreneurial without pretending every prototype is a company. It can give people ownership while maintaining enough shared structure that their work survives a graduation, a deadline, or a change of direction.

That means treating research operations as design work. How do we decide what to build? Who owns a component? What evidence makes a claim ready? How does a student learn to review, release, teach, and explain, not only implement? What should become shared infrastructure, and what should remain a focused experiment?

Who is here

The center now has research faculty, engineers, and a dozen PhD students, working with partners at Argonne, The HDF Group, Sandia, Livermore, Oak Ridge, Pacific Northwest, Utah, and Chicago.

Two outcomes I am especially glad about. Luke Logan and Jaime Cernuda, who both did their PhDs inside these systems, chose to stay on as research faculty in January 2026. Luke now leads software engineering, including day-to-day CLIO Core engineering; Jaime leads student development and day-to-day Agent and Relay development. Kun Feng runs the center’s infrastructure, and Eneko Gonzalez led its product work until September 2026.

What we are building

The visible outputs are systems, papers, talks, tutorials, and open repositories: IOWarp and CLIO, ChronoLog, LABIOS, the DeepIO line, and the completed Hermes and DTIO. Underneath them is an operating practice: regular technical exchange, explicit ownership, public artifacts when work is ready, and a willingness to connect foundational research with tools people can run. In 2025 I started a Friday seminar on AI-assisted programming for the graduate researchers, because the tools were changing faster than any curriculum.

My role

My role is to frame research directions, create proposals, build teams and partnerships, align milestones, integrate the systems, coauthor the research, report to sponsors, and remove obstacles without taking technical leadership away from other people. I do not always get the balance right. Too little structure strands people; too much structure turns research into ticket processing.

The goal is independent judgment. A strong mentee should eventually disagree with my architecture for reasons I did not see, then have the evidence and communication skill to carry a better idea forward.

The entrepreneurial edge

I use “entrepreneurial” here in its oldest useful sense: noticing an unmet need, assembling people and resources, accepting uncertainty, and making something durable enough to test in the world. Sometimes that becomes software infrastructure, a new collaboration, a community, or a proposal. The measure is whether the work creates new agency for the people around it.

What I am still learning

  • How to preserve speed without building around a single person’s memory.
  • How to make evidence and disclosure boundaries part of everyday development.
  • How to give emerging researchers ownership with a safe place to fail.
  • How to keep the center technically ambitious, publicly useful, and human as it grows.

Code, releases, papers, and records

  1. institution

    Gnosis Research Center

    The institutional home for the center's people, complete project catalog, and public resources.

  2. code

    GRC on GitHub

    Public software produced by the people and projects around the center.

  3. record

    Anthony Kougkas at Illinois Tech

    The institutional directory entry.

  4. record

    GRC team

    Current members and alumni of the center.

  5. record

    GRC research projects

    The funded projects, each with its own record.

  6. record

    GRC publications

    The center's bibliography, one page per paper.

  7. code

    GRC on Hugging Face

    The center's organization page on Hugging Face.

  8. press

    Introducing the Gnosis Research Center

    The center's announcement of October 25, 2023.

People

  • Xian-He Sun Co-founder and founding director
  • Luke Logan Assistant Research Professor; Head of Software Engineering
  • Jaime Cernuda Assistant Research Professor; Head of Student Development
  • Xiaoyang Lu Assistant Research Professor
  • Kun Feng Research Software Engineer; Head of Infrastructure
  • Eneko Gonzalez Research Software Engineer; Head of Product

Related notes

  1. essay

    Co-founding a research center

    The Scalable Computing Software lab at Illinois Tech is now the Gnosis Research Center. Why a lab with two decades of history became a center, what changes in practice, and what I am unsure about.

    6 min
  2. field note

    Vibe Coding Fridays

    Twelve Fridays, three hours each, teaching a research group to work with coding agents: the frame that stuck, the bake-off at week five, the checklist, and what I would change.

    6 min
  3. field note

    Reflecting on SC25

    What I noticed at SC25 in St. Louis, what the group brought, what a year of running the PDSW program committee taught me about how a workshop gets made, and a few notes for people attending their first SC.

    8 min
  4. field note

    Two PhDs, one architecture

    Luke Logan and Jaime Cernuda defended in the same semester. Their dissertations are the two halves of the runtime and the agent layer we are building, and they got there by owning the problems, not by executing mine.

    4 min
  5. essay

    A researcher's guide to funding proposals

    What I look for on a review panel, and what I try to do when I am the one writing. The reviewer reads for the question first, the budget is a claim about the plan, and a team is an argument.

    8 min
  6. personal

    2025 reflections

    A personal record of 2025: a promotion, a keynote, two dissertations, a summer of teaching, a loss, and the argument I kept making all year about what a researcher is for now.

    7 min
  7. essay

    Three pillars for a university in the AI era

    Faculty fluency first, a multi-vendor institutional strategy next, sovereign AI infrastructure as the destination. The case was made to the Illinois Tech faculty council and holds for any research university.

    7 min
  8. essay

    High-performance agentic computing

    The case for naming a discipline at the meeting point of HPC, autonomous AI, and domain science, with the multiplicative axiom behind it and my own doubt about whether it is a field or a decade-long transition.

    8 min
  • 2025–present

    CLIO

    A context layer for scientific AI agents and workflows.

  • 2024–present

    IOWarp

    Context management infrastructure for scientific AI.