Meetup Playlist
The Materials Project, a public initiative supported by funding from the US Department of Energy (DOE), computes properties of all known materials and remove guesswork from materials design. Discovery of novel materials can be targeted to the most promising compounds by screening and applying machine learning algorithms on MP’s computational data sets. Over 300,000 registered users from industry and academia are heavily using MP’s popular online gateway and API. Transitioning MP to fully operate in the cloud has hence become a cornerstone in the longevity of MP.
In this talk, we will present how MP runs a microservices-based network architecture on the AWS cloud and uses Datadog’s cloud monitoring services for end-to-end visibility into its elastic cloud resources. As MP moves its data products into the AWS Open Data Sponsorship Program, it is also building the foundation for gradually transitioning its data processing pipelines and the required computing to the cloud. The new cloud architecture allows MP to meet modern requirements for high availability, rapid solutions, increased security and scalability. The small core team implements infrastructure-as-code from the ground up to efficiently use its limited cloud computing and human resources.
Patrick holds M.Sc. and B.Sc. degrees (Dipl. Phys.) in Physics from the Technical University of Munich (Germany) and obtained his Ph.D. (Dr. phil. nat.) in high-energy nuclear physics from the Goethe-University in Frankfurt based on his work as a member of the large international STAR collaboration at the Relativistic Heavy Ion Collider at Brookhaven National Laboratory. Since 2014, he's been a core member of the Materials Project (https://materialsproject.org) at Lawrence Berkeley National Laboratory and is now the Senior Computer Systems Engineer on the project in charge of its Cloud & Scientific Data Architecture.