Publications

2026

1. WORKS

   

   The Other Half of Workflow Portability: Constructing Evidence-Backed HPC Site Profiles with Agentic Discovery

   Md Saiful Islam and Douglas Thain

   In 21st Workshop on Workflows in Support of Large-Scale Science (WORKS), SC26 Workshops, 2026

   Bib PDF Code

   @inproceedings{islam2026siteprofiles,
     title = {The Other Half of Workflow Portability: Constructing Evidence-Backed HPC Site Profiles with Agentic Discovery},
     author = {Islam, Md Saiful and Thain, Douglas},
     booktitle = {21st Workshop on Workflows in Support of Large-Scale Science (WORKS), SC26 Workshops},
     year = {2026},
   }

2. eScience

   

   Declare, Resolve, Reuse: A Portable Data Layer for Notebook Workflows Across HPC Systems

   Md Saiful Islam, Reyer Band, Jin Zhou, and 3 more authors

   In 2026 IEEE 22nd International Conference on eScience (eScience), 2026

   Bib HTML PDF

   @inproceedings{islam2026datalayer,
     title = {Declare, Resolve, Reuse: A Portable Data Layer for Notebook Workflows Across HPC Systems},
     author = {Islam, Md Saiful and Band, Reyer and Zhou, Jin and Malik, Tanu and Lannon, Kevin and Thain, Douglas},
     booktitle = {2026 IEEE 22nd International Conference on eScience (eScience)},
     year = {2026},
   }

3. eScience

   Portable By Design: Deploying Notebook-based Scientific Workflows across HPC Clusters

   Md Saiful Islam, Talha Azaz, Raza Ahmad, and 6 more authors

   In 2026 IEEE 22nd International Conference on eScience (eScience), Poster Paper, 2026

   Abs Bib HTML

   Notebooks are the default interface for exploratory science, but scaling them into distributed HPC workflows remains hard: provisioning and data staging are usually hand-coded into notebook cells, so the same notebook rarely runs unmodified elsewhere. We present an ecosystem that externalizes these concerns instead of burying them in code. TaskVine supplies scalable, fault-tolerant distributed execution. A backpack captures a notebook’s software, data, and compute requirements as a single portable specification. Floability Audit observes a workflow’s execution and automatically turns it into a backpack. Floability then provisions the backpack unmodified across heterogeneous HPC sites, resolving and caching data through a declarative data layer. Together, these pieces make portability a property of the ecosystem rather than a discipline imposed on the notebook.

   @inproceedings{islam2026portable,
     title = {Portable By Design: Deploying Notebook-based Scientific Workflows across HPC Clusters},
     author = {Islam, Md Saiful and Azaz, Talha and Ahmad, Raza and Hossain, A S M Shahadat and Baig, Furqan and Wang, Shaowen and Lannon, Kevin and Malik, Tanu and Thain, Douglas},
     booktitle = {2026 IEEE 22nd International Conference on eScience (eScience), Poster Paper},
     year = {2026},
   }

4. CCGrid

   Efficiently Reproducing Distributed Workflows in Notebook-based Systems

   Talha Azaz, Raza Ahmad, Md Saiful Islam, and 2 more authors

   In 2026 IEEE 26th International Symposium on Cluster, Cloud and Internet Computing (CCGrid), 2026

   DOI Bib

   @inproceedings{azaz2026reproducing,
     title = {Efficiently Reproducing Distributed Workflows in Notebook-based Systems},
     author = {Azaz, Talha and Ahmad, Raza and Islam, Md Saiful and Thain, Douglas and Malik, Tanu},
     booktitle = {2026 IEEE 26th International Symposium on Cluster, Cloud and Internet Computing (CCGrid)},
     pages = {74--83},
     year = {2026},
     doi = {10.1109/CCGrid68966.2026.00017},
   }

2025

1. eScience

   

   Backpacks for Notebooks: Enabling Containerized Notebook Workflows in Distributed Environments

   Md Saiful Islam, Talha Azaz, Raza Ahmad, and 6 more authors

   In 2025 IEEE International Conference on eScience (eScience), 2025

   Bib HTML PDF

   @inproceedings{islam2025backpacks,
     title = {Backpacks for Notebooks: Enabling Containerized Notebook Workflows in Distributed Environments},
     author = {Islam, Md Saiful and Azaz, Talha and Ahmad, Raza and Hossain, A S M Shahadat and Baig, Furqan and Wang, Shaowen and Lannon, Kevin and Malik, Tanu and Thain, Douglas},
     booktitle = {2025 IEEE International Conference on eScience (eScience)},
     pages = {169--177},
     year = {2025},
   }

2024

1. WORKS

   

   Shepherd: Seamless Integration of Service Workflows into Task-Based Workflows through Log Monitoring

   Md Saiful Islam and Douglas Thain

   In 19th Workshop on Workflows in Support of Large-Scale Science (WORKS), SC24 Workshops, 2024

   Bib PDF Code

   @inproceedings{islam2024shepherd,
     title = {Shepherd: Seamless Integration of Service Workflows into Task-Based Workflows through Log Monitoring},
     author = {Islam, Md Saiful and Thain, Douglas},
     booktitle = {19th Workshop on Workflows in Support of Large-Scale Science (WORKS), SC24 Workshops},
     pages = {2080--2087},
     year = {2024},
   }