Nico Christianson | Stanford Energy Fellow – Electric Grid Modernization

Duration: 2:44 447 transcript words EN summary

Video summary

Nico Christianson, a Stanford Energy Fellow who grew up in Pennsylvania and earned his PhD at Caltech, is dedicated to developing new algorithms and optimization methods that enable the safe use of artificial intelligence for accelerating the energy transition. His primary focus lies on creating tools capable of calibrating uncertainty within machine learning models, ensuring they can be reliably deployed across various scales—from small-scale energy resources to massive power grids—without compromising safety or stability. He chose Stanford specifically because it offers a unique environment where computer science and energy sciences intersect deeply, complemented by its proximity to the vibrant industrial ecosystem of the Bay Area. During his fellowship, Christianson has thrived within a community of passionate researchers from diverse fields such as biology and material science, gaining valuable insights into how computational approaches differ across disciplines while collaborating with individuals who share a commitment to driving progress in clean energy through rigorous research. Looking toward the future, Christianson will join Johns Hopkins University's Department of Computer Science as an assistant professor, where he plans to establish his own research group focused on bridging artificial intelligence with problems in clean energy systems. He remains eager to continue the collaborations forged at Stanford while expanding his work to ensure that AI tools not only improve efficiency and renewable integration but also come with robust reliability guarantees. Despite the immense promise of machine learning for optimizing existing resources and leveraging more renewable generation, Christianson emphasizes the critical risks associated with deploying models lacking safety assurances or requiring excessive energy for training. Consequently, his future research aims to develop innovative approaches that allow for efficient operation of energy systems while strictly maintaining model reliability and ensuring that the computational cost of training these advanced AI tools remains sustainable and effective.

Read the full video transcript

I'm Nico Christensen. I grew up in Media, Pennsylvania, and I did my PhD at Caltech. And during the fellowship, I'm working on developing new algorithms and optimization methods to help us reliably use AI and machine learning to help accelerate the energy transition. And specifically, I'm really interested in developing tools that can help us uh calibrate the uncertainty and confidence of these machine learning models so that we can know when we can safely deploy them to operate energy systems across scales, from [music] small-scale energy resources to large-scale power grids. I chose to come to Stanford for my postdoctoral fellowship because I knew I wanted to continue on in academia working on these slow-burn questions that require a lot of time and deep thought [music] to solve. And Stanford was the obvious choice because of the deep interdisciplinary ties between computer science and the energy sciences, as well as proximity to the hopping industrial ecosystem in the Bay Area. It's been really fantastic being in a community of similarly minded individuals across fields who are just really passionate about doing research [music] to drive the energy transition. And through the fellowship events, I get to learn from people in biology and material science and [music] working on things that I wouldn't otherwise know about. And it's also been really interesting to see how computational work [music] is approached in different fields, since I as a computer scientist can really learn from the problems in these other disciplines. After the fellowship, I'll be joining the Department of Computer Science at Johns Hopkins University [music] as an assistant professor, and I'm really looking forward to starting my own research group and [music] bridging artificial intelligence and problems in clean energy systems. And I'm also really looking forward to continuing the wonderful collaborations [music] that I've started during my time here at Stanford. When I think of the future, I think that AI and machine learning hold a lot of promise for making our energy systems more efficient, allowing us to take advantage of more renewable generation and operate existing resources more effectively. But, they also pose a lot of risk if we deploy AI models that don't have guarantees on safety or reliability. And they also pose risks due to the increasing energy demand that it requires to actually train these models in the first place. So, I'm very excited in the future to develop new approaches that can allow us to operate energy systems efficiently taking advantage of these AI [music] and machine learning tools, but also making sure that we have reliability guarantees and that these models are trained and deployed efficiently. >> [music]