Anthony Kougkas is an Associate Research Professor of Computer Science and the Co-Founder and Executive Director of the Gnosis Research Center at Illinois Institute of Technology, and Guest Research Faculty at Argonne National Laboratory. He serves as Co-Chair of AI and Research Infrastructure on IIT's Research Innovation Council. His work sits at the intersection of high-performance I/O, scientific data systems, AI infrastructure, and agentic software for research workflows. He builds runtimes, benchmarks, and AI tools that connect large-scale storage and scientific data formats with agents that inspect data, reason over evidence, interact with HPC systems, and preserve provenance. His recent work includes the IOWarp and CLIO ecosystem for context-aware scientific computing, with CLIO Agent for autonomous data management, CLIO Kit for MCP-based HPC and data tools, and CLIO Researcher for reproducible I/O experimentation; LABIOS for label-based agent I/O; and agentic survey infrastructure for gathering community requirements. Earlier he architected Hermes, ChronoLog, and DTIO. He has published more than 60 peer-reviewed papers, is an inventor on a U.S. patent on label-based data representation for I/O, and received the Karsten Schwan Best Paper Award (HPDC 2019) and the CCGrid Best Paper Award (2021). The MLPerf Storage benchmark (MLCommons) is built on DLIO, which he co-authored. He was General Vice Chair of SSDBM 2025 and Program Co-Chair of PDSW at SC (2024-2025), with TPC service at HPDC, IPDPS, CLUSTER, CCGrid, and SC.
86Publications
1Patent
12Projects
7Awards
15Grants
12Talks
11PhD advisees
Research vision
We are swimming in data, yet starving for context. Modern scientific and AI workflows generate data at rates that overwhelm conventional storage stacks, but the real bottleneck is not throughput. It is meaning. Current storage systems provide bytes, offsets, and file descriptors. AI agents need context, semantics, and provenance. Kougkas's research builds the translation layer between these two worlds through context engineering, a discipline of designing systems that orchestrate data, metadata, telemetry, and computational state to give autonomous agents the precise context they need. His systems (IOWarp, Hermes, LABIOS, ChronoLog) translate memory-centric theory into software developed with national laboratories; the MLPerf Storage benchmark is built on DLIO. The frontier is agentic AI for storage: systems that observe workload behavior, reason about data placement, and adapt autonomously.
Top collaborators
Featured publications
2019Karsten Schwan Best Paper Award
A. Kougkas, H. Devarajan, J. Lofstead, X.-H. SunThe 28th International Symposium on High-Performance Parallel and Distributed Computing (HPDC'19), Phoenix, USA 2019. pp. 13-24. Karsten Schwan Best Paper Award
LABIOS introduced label-based I/O, with asynchronous data movement giving up to 17x higher I/O performance in the HPDC'19 evaluation. Karsten Schwan Best Paper Award at HPDC 2019.
2018
A. Kougkas, H. Devarajan, X.-H. SunThe 27th ACM International Symposium on High-Performance Parallel and Distributed Computing (HPDC), Tempe, AZ, USA, 2018. pp. 219-230
Hermes provides multi-tiered I/O buffering that places data across memory, NVMe, SSD, and parallel file systems. The predecessor of IOWarp.
2021Best Paper Award (First Prize)
H. Devarajan, H. Zheng, A. Kougkas, X.-H. Sun, V. VishwanathThe 2021 IEEE/ACM International Symposium in Cluster, Cloud, and Internet Computing (CCGrid'21), May 17 - 20, 2021 Best paper award
DLIO is a data-centric benchmark for deep learning I/O. The MLPerf Storage benchmark (MLCommons) is built on it. Best Paper at CCGrid 2021.
2020
A. Kougkas, H. Devarajan, K. Bateman, J. Cernuda, N. Rajesh, X.-H. SunThe 36th International Conference on Massive Storage Systems and Technology (MSST'20), Oct. 29-30, 2020
ChronoLog captures distributed activity data with temporal ordering, for provenance tracking and replay across scientific workflows.
2020
A. Kougkas, H. Devarajan, X.-H. SunACM Transactions on Storage (TOS'20), Vol 16, No 4, Article 22, Nov. 2020
A journal article on storage bridging for converging HPC and big data storage systems through unified data management middleware.
2018
A. Kougkas, H. Devarajan, X.-H. Sun, J. LofsteadThe IEEE International Conference on Cluster Computing 2018 (Cluster'18), Belfast, UK2018. pp. 290-301.
Harmonia introduced a conflict-free coordination mechanism for concurrent I/O operations in multi-tiered buffering systems, enabling lock-free data movement.
Browse all 86 publications →Research interests
Agentic AI for Systems
Context Engineering
Parallel and Distributed I/O
HPC Storage Systems
Intelligent I/O Buffering
Storage Semantics and Data Labels
Deep Learning I/O
Computational Storage
Data-Centric Middleware
Scientific Workflow Optimization
Projects
A distributed shared log that orders activity and provenance data with physical time and moves records across memory, local storage, and archival tiers.
An active-storage framework that computes derived quantities in transit and queries enriched metadata. The Hades paper reports 3 to 4 times faster analysis on tested Gray-Scott workflows.
Data-path methods for scientific AI, from the DLIO benchmark and Viper model transfer framework to UnboxKV characterization and PKAS scheduling for KV caches under concurrent LLM inference.
A task-based I/O runtime developed with Argonne National Laboratory to connect HPC, Big Data, and ML data stacks. The SSDBM'25 paper reports 49.6% better I/O performance for online translation with DataTask caching than for offline translation.
Distributed I/O buffering middleware that coordinates data placement across DRAM, NVMe, burst buffers, and parallel file systems.
IOWarp develops CLIO: scientific data and coding agents, shared tools, and a context runtime connecting AI-assisted research to real data and systems.
Unified data access middleware that translates between parallel file-system and object-store semantics.
LABIOS represents I/O as labels that carry an operation, data pointer, and routing metadata. Its public prototype routes labels from applications and agents through a dispatcher to Databot and Pipeline workers; the Agentic tier remains planned.
A DOE Genesis Mission project led by Pacific Northwest National Laboratory that ports scientific workflows to HPC systems and tunes where their tasks and data run. GRC contributes workflow performance research and CLIO.
A DOE Genesis Mission project led by Argonne National Laboratory that captures provenance across scientific workflows and traces and attributes threats to their data and results. GRC contributes CLIO from IOWarp.
StoreHub combined two community instruments, a national workshop, and a landscape analysis to assess the need for dedicated data storage research infrastructure. Its public report recommends a staged federation built on existing facilities.
WisIO analyzes HPC I/O traces through time, process, and file views, then connects metric-level bottlenecks with root causes.
PhD advisees
Current (9)
HPC, Distributed Storage, Metadata Management
HPC, GPU-Accelerated Storage, I/O Characterization
HPC, Distributed Storage, I/O Characterization
Distributed Storage, Parallel and Distributed Systems, LLM/DNN Inference Optimization
HPC, Distributed Storage, Task Systems
Parallel and Distributed Systems, HPC, ML-Assisted Systems
HPC
AI-Driven Scientific Workflows, Model Context Protocol (MCP) for HPC, Agentic Systems for Scientific Computing
HPC, Systems
Graduated (2)
Jaime Cernuda Garcia2026FlowForge: Taming the Log-Data Storm in Distributed ComputingAssistant Research Professor, Illinois Institute of Technology Luke Logan2025RECOMBINANT: A Composable, Extensible, Intelligent I/O Platform for Integrated High-Performance I/O ManagementAssistant Research Professor, Illinois Institute of Technology Awards and honors
2020s
2023ACM SIGHPC Certificate of Appreciation, ACM SIGHPCService
2021Best Paper Award (First Prize), CCGrid'21, IEEE/ACM CCGridBest paper
2010s
2019Karsten Schwan Best Paper Award, HPDC'19, ACM HPDCBest paper
2018Best Research Assistant, Illinois Institute of Technology, Department of Computer ScienceInstitutional
2014Tsao Fund Scholarship for Outstanding Academic Performance, Illinois Institute of TechnologyInstitutional
2014Best Teaching Assistant, Illinois Institute of Technology, Department of Computer ScienceInstitutional
2012Fellowship for Academic Excellence, University of PiraeusInstitutional
Keynotes and invited talks
“From Storage Optimization to Context Orchestration: Context Engineering for Environmental Data Systems”
Invited talkESIP Monthly Meeting, MCP for Environmental Data SessionVirtualFeb 2026
“From Storage Optimization to Context Orchestration: Why Your AI Infrastructure is Already Obsolete”
Invited talkCS695 Doctoral Seminar, Illinois Institute of TechnologyChicago, ILFeb 2026
“Data Management in the Age of AI: Bridging Scientific Workflows with Intelligent Systems”
Invited talkCSE 5449 Guest Lecture, Ohio State UniversityColumbus, OHNov 2025
“From Chatbots to Autonomous Agents: A Practical Guide to AI Coding Tools for HPC Researchers”
Invited talkInnovative Computing Laboratory (ICL) Friday Lunch Talk, University of Tennessee, KnoxvilleKnoxville, TNOct 2025
“Data Management for AI-Powered Science Agents: From Storage Optimization to Context”
Invited talkArgonne National Laboratory, MCS Division SeminarLemont, ILSep 2025
“AI Infrastructure for Scientific Computing”
KeynoteAI4Sys Workshop at ACM HPDC 2025Jun 2025
“Data Management for Genomics”
Invited talk7th IFSH High-Throughput Sequencing SymposiumChicago, ILFeb 2025
“AI4Science Research Initiatives”
Invited talkIIT University-Wide SeminarChicago, IL2025
“IOWarp and Storage Integration”
Invited talkHDF User Group Annual MeetingChicago, IL2024
3 earlier talks, 2021
“Hermes Multi-Tiered Storage”
Invited talkOak Ridge National Laboratory ATS Seminar SeriesVirtual2021
“Hermes and HDF5”
Invited talkHDF User Group Annual MeetingVirtual2021
“Multi-tiered Storage Systems”
Invited talkRex-IO Workshop on Multi-tiered StorageVirtual2021
Professional service
Program Co-ChairPDSW at SC2024-2025
General Vice ChairSSDBM International Conference2025
Web ChairIEEE eScience Conference2025
Local ChairHDF User Group Meeting2024
Technical Program CommitteeACM HPDC2022-2026
Technical Program CommitteeIEEE IPDPS2023-2025
Technical Program CommitteeIEEE CLUSTER2018, 2021-2022, 2024
Technical Program CommitteeIEEE/ACM CCGrid2020, 2024
TPC / Data Analytics Committee / External ReviewerACM/IEEE SC Conference Series2019-2025
ReviewerIEEE Transactions on Parallel and Distributed Systems (TPDS)2023-present
ReviewerIEEE Transactions on Computers (TC)2024-present
ReviewerIEEE Transactions on Networking (ToN)2022-present
ReviewerACM Computing Surveys2024-present
Review Panel ExpertDOE ASCR (Advanced Scientific Computing Research)2024
Co-Chair, AI and Research InfrastructureUniversity Research Innovation Council, Illinois Institute of Technology2025-present
MentorSC Mentor-Protege Program2021, 2022, 2024
4 earlier service roles, 2019 to 2021
Poster ChairICPP International Conference2021
Technical Program CommitteeIEEE MSST2019
Technical Program CommitteeWorkshops: HPC-IODC, ScienceCloud, FlexScience, HiPESvarious
MentorIPDPS PhD Forum2019, 2020
Funded research
Active
PIIntentIO: Machine Learning Driven Approach to Extract I/O Intents to Improve Efficiency of Deep Learning Training
WisIOCo-PICollaborative Research: Frameworks: IOWarp: Bending the I/O Fabric for Advancing AI-Infused Scientific Workflows (OAC-2411318)
IOWarpCo-PILABIOS: Storage Acceleration via Data Labeling and Asynchronous I/O (OAC-2313154)
LABIOSCo-PIIdentifying and Mitigating I/O Bottlenecks for HPC Workflows
PIDeepIO: Deep Learning Inference I/O Optimization (2026 renewal)
DeepIOPIOptimizing metadata for efficient discovery of inference models in Deep Learning workflows
CoeusCo-PIFrameworks: Collaborative Research: ChronoLog: A High-Performance Storage Infrastructure for Activity and Log Workloads (OAC-2104013)
ChronoLogCompleted
Co-PICollaborative Research: CIRC: Planning-C: StoreHub: A Community Infrastructure for Shaping the Future of Data Storage Research (CNS-2346504)
StoreHubCo-PIDTIO: Enabling Computational Storage using Data-Tasks and Asynchronous I/O (DE-SC0024593)
DTIOPIDeepIO: Deep Learning Inference I/O Optimization (3J-60004-0013A)
DeepIOCo-PICoeus: Accelerating Scientific Insights Using Enriched Metadata (DE-SC0023263)
CoeusCo-PIDFlow: Optimizing Data-Intensive Workflows by Understanding Data
PortusCo-PICSR: Small: IRIS: A Unified Data Access Framework for the Merging of Compute-Centric and Data-Centric Storage (CSR-1814872)
IRISPIDeepMeta: Optimizing Metadata for Efficient Discovery of Inference Models
DeepIOCo-PIFramework: Software: NSCI: Collaborative Research: Hermes: Extending the HDF Library to Support Intelligent I/O Buffering (OAC-1835764)
HermesNews mentions