Jackson Clark

hackson@uiuc:~

$ whois hackson

name: Jackson Clark
role: PhD Candidate, 3rd year
org: Siebel School of Computing & Data Science, UIUC
advisor: Tianyin Xu
building: SREGym
focus: AI for SRE, systems reliability, post-training & RL, evals
email: jclark58@illinois.edu

$ cat about.txt

I'm a third year PhD candidate in the Siebel School of Computing and Data Science at the University of Illinois Urbana-Champaign (UIUC). I am "watched" by my advisor Tianyin Xu.

I'm currently building SREGym to see if AI can resolve production incidents. SREGym is part of Laude Slingshots // TWO. See my interview with Andy Konwinski and a brief demo.

My research explores practical AI and agent techniques for safety-critical computer systems and infrastructures. My current work is in the AIOps space, specifically in AI for Site Reliability Engineering (SRE). Previously, I collaborated with Microsoft Research and IBM Research to build LLM benchmarks (AIOpsLab, ITBench) and agents (STRATUS) for SRE.

In the past, I worked at NASA in the Space Technology Mission Directorate, where I helped develop space systems for the International Space Station. I'm also an instructor at the Discovery Partners Institute at the University of Illinois, where I teach a professional development class on Data Science to CPS high school teachers.

$ cat links.txt

✉ email ◈ scholar ⌥ github ⬡ linkedin ◎ openreview ≡ dblp ↓ cv

$ tail -n 10 news.log

2026-10NEWStarted an internship at Meta Superintelligence Labs, working on coding evaluations with Ofir Press.
2026-09PAPERSREGym accepted to NeurIPS 2026. See you in Sydney! 🇦🇺
2026-06NEWJoined Bespoke Labs to build RL environments for SRE.
2026-05PASSPassed my PhD qualifying exam.
2026-05NEWReleased the SREGym paper on arXiv.
2026-04PAPERSREGym accepted to the CAIS 2026 Demo Track.
2025-10GRANTAwarded a Laude Institute Slingshot grant for SREGym.
2025-09PAPERSTRATUS accepted to NeurIPS 2025.
2025-05ORALITBench selected for Oral Presentation (Top 1%) at ICML 2025.
2024-08INITStarted my PhD at the Siebel School of Computing and Data Science, UIUC.

$ ls -lt ./publications/

NeurIPS 2026 · Evaluations & Datasets

SREGym: A Live Benchmark for AI SRE Agents with High-Fidelity Failure Scenarios

Jackson Clark*, Yiming Su*, Saad Mohammad Rafid Pial, Yifang Tian, Lily Gniedziejko, Hans-Arno Jacobsen, Yinfang Chen, Tianyin Xu. * equal contribution

pdf site

bibtex

@article{clark:arxiv:26,
  author  = {Jackson Clark and Yiming Su and Saad Mohammad Rafid Pial and Yifang Tian and Lily Gniedziejko and Hans-Arno Jacobsen and Yinfang Chen and Tianyin Xu},
  title   = {{SREGym: A Live Benchmark for AI SRE Agents with High-Fidelity Failure Scenarios}},
  journal = {arXiv:2605.07161},
  year    = {2026},
  month   = may,
  eprint  = {2605.07161},
  archivePrefix = {arXiv}
}

CAIS 2026 · Demo
SREGym: A Live Training Ground for AI SRE Agents with High-Fidelity Failure Drills

Jackson Clark, Yiming Su, Saad Mohammad Rafid Pial, Lily Gniedziejko, Tianyin Xu.

pdf

NeurIPS 2025

STRATUS: A Multi-agent System for Autonomous Reliability Engineering of Modern Clouds

Yinfang Chen*, Jiaqi Pan*, Jackson Clark*, Yiming Su*, Noah Zheutlin, Bhavya Bhavya, Rohan Arora, Yu Deng, Saurabh Jha, Tianyin Xu. * primary author

pdf arxiv

bibtex

@inproceedings{chen2025stratus,
  title     = {{STRATUS}: A Multi-agent System for Autonomous Reliability Engineering of Modern Clouds},
  author    = {Chen, Yinfang and Pan, Jiaqi and Clark, Jackson and Su, Yiming and Zheutlin, Noah
               and Bhavya, Bhavya and Arora, Rohan and Deng, Yu and Jha, Saurabh and Xu, Tianyin},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year      = {2025},
  eprint    = {2506.02009},
  archivePrefix = {arXiv},
  primaryClass  = {cs.DC}
}

ICML 2025 ORAL · TOP 1%

ITBench: Evaluating AI Agents across Diverse Real-World IT Automation Tasks

Saurabh Jha, Rohan Arora, Yuji Watanabe, Takumi Yanagawa, Yinfang Chen, Jackson Clark et al. Vancouver, Canada, July 2025. Spotlight poster.

pdf arxiv

bibtex

@inproceedings{jha2025itbench,
  title     = {{ITBench}: Evaluating {AI} Agents across Diverse Real-World {IT} Automation Tasks},
  author    = {Jha, Saurabh and Arora, Rohan and Watanabe, Yuji and Yanagawa, Takumi
               and Chen, Yinfang and Clark, Jackson and others},
  booktitle = {Proceedings of the 42nd International Conference on Machine Learning (ICML)},
  year      = {2025},
  address   = {Vancouver, Canada},
  note      = {Oral Presentation (Top 1\%)}
}

$ cat invited_talks.txt

SREGym — Microsoft Research, System Intelligence Seminar. Spring 2026
STRATUS — University of Toronto, ECE1770. Hosted by Yifang Tian and Hans-Arno Jacobsen. Mar 2026

$ ls ./tutorials/

# I've been giving many tutorials on AIOps and SRE agent research.

An open platform for AI-driven systems research on cloud system management (ITBench Tutorial) — SOSP 2025
Benchmarking of AI Agents for IT Automation Tasks using ITBench — DSN 2025
AIOpsLab in Action: An Open Platform for AIOps Research — FSE 2025
Benchmarking of AI Agents for IT Automation Tasks using ITBench — IIDAI Workshop Spring 2025

$ cat mentees.txt

Saad Mohammad Rafid Pial — Research intern at UIUC (BSc, BUET). Working on AI for SRE and cloud reliability.
Lily Gniedziejko — Undergraduate, Computer Science, UIUC.

$ grep -i "grant" cv.pdf

Catalyst Grant from Cohere — 2026
Slingshot Grant from Laude Institute for SREGym — 2025

$ cat service.txt

OSDI 2025 Artifact Evaluation

$ history | grep job

Meta Superintelligence Labs · Research InternOct 2026 – now

Working on coding evaluations with Ofir Press.
Bespoke Labs · Building RL Environments for SREJun 2026 – now

Building reinforcement learning environments for Site Reliability Engineering.
xlab, UIUC · Graduate Research AssistantFall 2024 – now

Co-authored STRATUS, an LLM-based multi-agent system for autonomous SRE that formalizes a Transactional Non-Regression (TNR) safety specification and shows 1.5–5.4× improvement over state-of-the-art AI SRE agents on AIOpsLab and ITBench. Lead maintainer of AIOpsLab (700+ GitHub stars), co-author of ITBench with IBM. Build microservice testbeds, eBPF-based chaos engineering, Kubernetes workload orchestration, and LLM-driven fault diagnosis pipelines on Prometheus / Jaeger / OpenTelemetry stacks.
NASA Marshall · Software Engineer Intern, HOSCSummer 2023

Built a C++ forward error correction library (BCH, LDPC, convolutional codes) to CCSDS standards for the GPDM mission and contributed automation tools supporting flight controllers. Huntsville, AL
NASA Ames · Research Intern, Intelligent Systems DivisionSummer 2023

Led reinforcement learning research for multi-agent pathfinding (MAPF) in urban air traffic management using PyTorch / TensorFlow on a supercomputing cluster. Mountain View, CA (remote)
NASA Goddard · Software Engineer & Data Science Intern, HEASARCFall 2022 – Spring 2023

Trained ML models to classify neutron star spectra from NICER payload data; redesigned SkyView, a React + Flask catalog for querying astrophysics datasets by sky position. Greenbelt, MD (remote)
NASA Marshall · Software Engineer Intern, POICSummer 2022

Designed an automation tool for ISS flight controllers to streamline timeline change processes. Huntsville, AL
Freelance · Software EngineerSep 2020 – Sep 2023

Earned $60k+ on Upwork shipping cross-platform apps in Electron.js, React, and Express; managed full client lifecycle from proposal to delivery.
RoboThink · Software EngineerJun 2018 – Sep 2020

Lead developer for RoboThink's educational tools built with Electron.js and React.

$ cat teaching.txt

Instructor, Discovery Partners Institute, University of Illinois. Designed and taught a two-week Data Science Intensive for Chicago Public School teachers so they can build their own data science lessons. Summer 2024, 2025, 2026
Course Assistant, Data Science Discovery (CS 107), UIUC. Spring 2023 – Spring 2024

$ cat education.txt

Ph.D. Computer Science, University of Illinois Urbana-Champaign — 2024–?
B.S. Data Science, University of Illinois Urbana-Champaign — 2020–2024

$ ls ./press/

SREGym/

STRATUS/