Skip to content Skip to navigation

Openings

Post-doctoral positions

The Martinez group is recruiting. While we do a broad range of research in traditional theoretical chemistry, a couple of our projects have developed a life of their own and require specialist development.

  • Machine Learning and Artificial Intelligence for Retrosynthesis

  • Interactive Molecular Dynamics

  • Ab initio nanoreactor 

 

Machine Learning and Artificial Intelligence for Retrosynthesis

Postdoctoral Position available in the Martinez lab at Stanford University, in strong collaboration with Alan Aspuru-Guzik (University of Toronto) and the Vector Institute

The Martinez and Aspuru-Guzik labs are looking for an enthusiastic and experienced postdoctoral researcher to carry out research at the interface of Artificial Intelligence and Machine Learning (ML) and Chemistry. In particular, the postdoctoral researcher will be developing and applying new ML algorithms for the problem of retrosynthesis, which is the task of finding the reagents and reactions that are more suitable to make a compound in the lab. The candidate will be working with the teams at both locations and co-advised by the two Principal Investigators.

A PhD in Computer Science, Computational Chemistry, Physics, Materials Science or related field is required.

Knowledge of chemistry, if the candidate is a computer scientists is not required, but would be welcome. Knowledge of basic machine learning tools such as deep networks is required if the postdoc has a chemistry or materials science background.

More information can be found in the research group websites: mtzweb.stanford.edu and matter.toronto.edu

Interested applicants should contact Dr. Todd Martinez at: Todd.Martinez@stanford.edu cc. Dr. Alan Aspuru-Guzik at aspuru@utoronto.ca

 

Interactive Molecular Dynamics

For small to medium size molecules, we can perform several ab initio calculations per second. This opens up the possibility of real-time interactions with a fully quantum simulation of a molecule. We have all become accustomed to seeing animations of molecular dynamics trajectories, but usually these have been generated over days and weeks. Using the Unity game engine we have built a prototype interactive ab initio molecular dynamics system with a haptic controller, allowing the user to grab atoms and drag them around. The next step is to extend this to virtual and augmented reality interfaces, and to scale up the calculations to mixed QM/MM models so we can work with larger molecules and molecules in explicit solvent. There are myriad potential applications and some fascinating problems in user interface design to be solved.

For this project we are looking for people who are excited by the possibilities of interacting with molecules. The skills you’ll need are more game design than quantum chemistry. It would help if you had experience with modelling molecules in some way, but it’s more important you have imagination and drive. This is definitely a programming job. Unity is built in C# but if you’ve worked with python or C++ in the past you’ll be fine. If your only programming experience is Fortran you might be in for a steep learning curve, which is fine by us so long as you’re committed to the climb. Down the track we plan to make use of the Unity ML Agents framework so experience with machine learning would definitely be a plus. 

Please send a brief email to keiran@stanford.edu with a CV and cover letter.  Tell us what excites you about these projects and where you would like to take them.  If you have projects on Github or other open platforms, please provide a link. 

 

Ab Initio Nanoreactor

The ab initio nanoreactor is the obvious thing to do when you can solve the time independent schrodinger equation for 300 atoms quickly enough to run classical molecular dynamics for nanoseconds. It’s like an in silico test tube. Some of the problems we’ve encountered are seeing the same elementary reactions too often and not finding enough interesting reactions. This is now a sophisticated code base with the potential to transform our understanding of reactive chemistry. The raw trajectories are processed into a graph of chemical reactions on which we can perform various forms of machine learning, such as path finding and link completion. Paths in this are synthesis plans and missing links are reactions that should be possible but have not yet been observed.

To further this research project we need people with modern software development experience and an interest in machine learning. The code base is Python, with some C++ for performance critical components. The web technologies involved are Tornado, Celery, Mongo, and of course Docker and Kubernetes. The ML layer is built in Keras and pyTorch. An interest in chemistry is advisable, although most of the work will be software development.  

Please send a brief email to keiran@stanford.edu with a CV and cover letter. Tell us what excites you about these projects and where you would like to take them.  If you have projects on Github or other open platforms, please provide a link.

 

 

Stanford is an equal employment opportunity and affirmative action employer. All qualified applicants will receive consideration for employment without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, disability, protected veteran status, or any other characteristic protected by law. Stanford also welcomes applications from others who would bring additional dimensions to the University’s research, teaching and clinical missions.