Sanjeev Raja

I’m a third-year Computer Science PhD student at Berkeley AI Research (BAIR) advised by Aditi Krishnapriyan.

My interests lie at the intersection of machine learning and atomistic modeling, with a particular focus on developing ML methods to accelerate molecular dynamics simulations. Recently, I have also been working on generative modeling of molecular conformational ensembles.

In the past, I worked on ML for accelerating climate simulations. I worked with Jaideep Pathak and Anima Anandkumar on FourCastNet, which was a state-of-the-art weather prediction model at the time.

I love Carnatic music, and I am a semi-professional mridangam player. I developed a state-of-the-art Carnatic raga classifier capable of identifying 150 common Carnatic scales. I also enjoy tennis, and particularly Roger Federer. Unfortunately, however, no amount of machine learning will reverse the result of the 2019 Wimbledon Final.

news

Jun 1, 2025	Joining Valence Labs in Montreal for a research internship.
May 1, 2025	Our paper on repurposing pretrained atomistic generative models for transition path sampling has been accepted at ICML 2025.
Jan 10, 2025	Check out our paper (accepted at ICLR 2025) and talk on distilling machine learning force field foundation models into fast models specialized for particular subsets of chemical space.
Feb 21, 2024	Check out our paper (accepted at TMLR) and talk introducing StABlE, a multi-modal approach for training stable machine learning force fields that combines supervision from quantum-mechanical energies/forces and system observables!