Group Leader Tapped for New Advanced Data and Workflow Group
Sukumar to fill NCCS, OLCF role
Dr. Sreenivas Rangan Sukumar has been named the group leader for the recently established Advanced Data and Workflow Group in the National Center for Computational Sciences (NCCS) at the US Department of Energy’s (DOE’s) Oak Ridge National Laboratory (ORNL). Sukumar assumed his new duties on October 1.
The group will serve as the interface between the Oak Ridge Leadership Computing Facility (OLCF), a DOE Office of Science User Facility located at ORNL, and the Compute and Data Environment for Science (CADES). As group leader, Sukumar will be the liaison between the two organizations. This group will design, build, and offer creative data-science workflow solutions to enable interactive data-driven discovery for DOE-focused science that require scale and performance on leadership compute architectures hosted by the OLCF and big data architectures hosted by CADES at ORNL.
“I am very enthusiastic about this new group,” Sukumar said. “Even in this decade of big-data frenzy, there aren’t many data science jobs that can make a difference across science domains spanning astronomy to zoology. It is exciting to be part of a fantastic team of data scientists, developers, and visualization experts that share a similar passion for science and are driven by the opportunity to apply their skills on problems for the greater social good while tackling tough computer science challenges at scale.”
Sukumar received his PhD in electrical engineering from the University of Tennessee in 2008. After completing a post-doctoral assignment at ORNL, he joined the Computational Science and Engineering Division as a research staff member. He served most recently as a member of the Computational Data Analytics Group and as a data scientist in residence for the ORNL Health Data Sciences Institute.
His vision for the new group is threefold, focusing on short-term, mid-term, and long-term goals.
“In the short term we’ll prioritize workflow requirements of OLCF scientific users to provide off-the-shelf analytical, visualization, and machine learning tools. This exercise with several pilot projects across domains will reveal specific challenges and gaps posed by big data in science compared to problems in industry applications,” Sukumar said.
“As data pilots mature, future workflows will demand a combination of heterogeneous architectures, such as Titan, Rhea, and EOS, and CADES resources such as Urika-XA and GD, with higher expectations of improved performance. A scientific user could say, ‘I have this workflow. The results during our pilot took 20 minutes. It would be beneficial if we can get our answer in 20 seconds.’ We’ll help the user toward that goal.”
“In the mid-term, the group will develop expertise to understand scalability of algorithms on a variety of current and emerging architectures. We’ll do so by benchmarking on both CADES and the OLCF hardware,” Sukumar said.
In the long term, he hopes to build a collection of tools—an app store, of sorts—that will scale on different architectures with portable open-source data science codes. “We will be solving DOE-specific science problems and publishing our software products and artifacts as mini-apps for big data that can be applied on problems beyond the OLCF user community.”
Other priorities identified by Sukumar include assisting with data-related needs for principal investigators submitting proposals for the Innovative and Novel Computational Impact on Theory and Experiment (INCITE) projects next year. Also, Sukumar said the hiring (and retention) of staff members using a strategic sustainable plan particularly for data science and visualization positions is critical to the success of this group.
With those objectives in place, Sukumar said he is excited about the upcoming responsibilities.
“Science is becoming more and more data driven,” he said. “And this group’s charter is to leverage leadership-class computing resources to enable fundamental scientific discoveries that are going to change the world in the future. While doing the best with what we already have, we are also going to be thinking ahead on what will science workflows look like on the exascale machine, what problems can be solved then that we cannot solve today, and how can we help our science users solve such problems? I’m looking forward to doing that.”
Oak Ridge National Laboratory is supported by the US Department of Energy’s Office of Science. The single largest supporter of basic research in the physical sciences in the United States, the Office of Science is working to address some of the most pressing challenges of our time. For more information, please visit science.energy.gov.