Poised between the University of Pittsburgh’s health care AI innovators and Carnegie Mellon University’s machine learning robots is the Pittsburgh Supercomputing Center (PSC), a decades-long joint project of the two universities that is helping to drive Pittsburgh as an AI ecosystem.
“It's a lot of fun,” says Barr von Oehsen, director of PSC. “Our relationship between Pitt and Carnegie Mellon is unique because instead of just one, we’re associated with two very strong research programs. Based on that relationship, we can build systems locally that we can then turn into national solutions.”
But how can people at different institutions collaborate while maintaining secure systems? Von Oehsen has led efforts to build networks using the authentication at a researcher’s home institution to be accepted at other institutions to access computing, data and research.
“We are trying to integrate systems that are currently in silos into an open but secure national platform where PSC is the entry point to that infrastructure.”
“We looked at ways to build infrastructure that makes it easier for people to collaborate without having to jump through security hoops,” von Oehsen explains. “We are trying to integrate systems that are currently in silos into an open but secure national platform where PSC is the entry point to that infrastructure.”
Von Oehsen points out that with the demand of AI, many universities are now struggling with the power and cooling needs that AI systems require. He says that now is the time for research institutions to come together to develop common integrated solutions that will spur collaboration, increase innovation and save money. For Pittsburgh, PSC is part of that solution.