Although ~200 organohalide hybrid perovskites structures are known, there is no theoretical understanding that can be used to predict structure, reaction conditions, or the effect of structure on long-term stability. The current discovery process for new organohalide perovskites relies on students performing individual experiments in a slow, trial-and-error way. There is no underlying physical or empirical theory relating reaction "inputs" and crystal quality or structural "outputs". Our goal is to replace this with a largely autonomous materials discovery system combining active machine learning techniques (software) with a robotic synthesis platform (hardware) operating across multiple geospatially separated laboratories.
The most important metrics of success of this project will be the ability to generate new materials that conform to the given structural/property specifications along with an accompanying explanatory theory describing the relationships between reagent properties, reaction conditions, and the resulting structure/property.
We're hiring for three positions towards this effort, and the full team will be composed of these three positions along with three faculty (Sorelle Friedler, Alexander Norquist, and Joshua Schrier), an experimental chemistry postdoctoral researcher, and undergraduate researchers and programmers. All members of the team will also interact and coordinate with portions of the experimental team located in San Francisco and Berkeley as well as coordinated groups engaged in a similar research effort across the country.
The deadline for all positions is rolling. We expect to start work on the project in late Feburary 2018.
The application due dates listed at the HR links below are wrong and should be ignored. Applications will be considered as they are received with the goal of hiring soon.
Cheminformatics Postdoctoral Researcher
The hired postdoctoral research fellow will conduct scientific research on organohalide perovskite synthesis. This work includes: developing and performing bench-scale feasibility tests of new chemical reactions for organohalide perovskite syntheses suitable for automated reactions; working with remote sites to implement these synthetic routes; planning experiments to be performed at remote sites; work closely with a cheminformatics and computer science postdoctoral fellows to test new types of machine learning tools for experiment-planning.
The successful candidate will have previous, doctoral-level experiences in synthetic solid-state or inorganic chemistry or a closely related field. Previous experience in organohalide perovskite synthesis is highly desirable.
Computer Science Postdoctoral Researcher
The hired postdoctoral research fellow will conduct research on applications of active learning and interpretable models to predictive chemistry applications. This work focuses on fundamental research in interpretable models, especially as applied to active learning, along with the construction and evaluation of machine learning models for predictive experimental chemistry. It will include experimental analysis of these models as part of a recommendation system informed by a chemistry expert. The postdoctoral fellow will work closely with a cheminformatics postdoctoral fellow, chemistry postdoctoral fellow, and software engineer to implement new types of experimental planning tools, with a focus on automating the development and testing of chemical hypotheses.
The successful candidate will have previous, doctoral-level experiences in computer science, ideally in the area of machine learning or a related field. Background or interest in chemistry, computer vision, and/or computer graphics considered a plus.
The hired software engineer will develop open-source code supporting scientific research on the application of machining learning methods and simulations to problems in materials science. This work involves: database development and maintenance, web-portal development and maintenance, and other software development and documentation tasks. The software engineer will work closely with postdoctoral researchers from computer science and chemistry to turn research code into production-quality software.
The successful candidate will have previously demonstrated experience in Python, Django, and MySQL development. Previous experience with database and API design is highly desirable. Interest in or experience with machine learning, computer vision, and/or computer graphics considered a plus.