• Oct 2023: New paper at MICCAI: InverseSR: 3D Brain MRI Super-Resolution Using a Latent Diffusion Model. Updated code here.
  • Feb 2023: Talk about my research in the UCSC AI club.
  • July 2022: I officially started as an Assistant Professor in the Computer Science and Engineering department at UC Santa Cruz! I am looking to build my lab and hire PhD students and postdocs.
  • Dec 2021: Exciting news! Our Bayesian Image Reconstruction (BRGM) paper got the Best paper award at the NeurIPS Deep Generative Models and Downstream Applications Workshop.
  • Nov 2021: Our Bayesian Image Reconstruction using Deep Generative Models paper has been accepted for an oral presentation at the NeurIPS Deep Generative Models Workshop. See recorded talk and slides.
  • Oct 2021: Our TADPOLE Challenge results paper has been accepted in the MELBA journal.
  • June 2021: New extension to BRGM: sampling of multiple reconstruction through Variational Inference. Updated paper here. Also, new updated code-implementation in PyTorch here, for both Bayesian-MAP and VI inference schemes.
  • January 2021: New pre-print: Bayesian Image Reconstruction using Deep Generative Models (BRGM). I used StyleGAN2 for reconstructing images corrupted by any corruption process. SOTA performance on super-resolution, at low input-resolution, and inpainting. Code and pre-trained models now available on the project page.
  • September 2020: Talk on GANs at the Harvard DBMS Clinical Lecture Series. See video recording and slides.
  • July 2020: Talk at AAIC conference on final results of the TADPOLE Challenge.
  • July 2020: Poster at AAIC conference on BrainPainter.
  • June 2020: Wrote blog post on why generative models are important
  • February 2020: TADPOLE Challenge results paper now on arxiv. We present all participant algorithms and complete analysis of the competition results.
  • September 2019: Three of my papers were accepted at MICCAI 2019!
    • TADPOLE (oral at PRIME workshop): preliminary results of Alzheimer’s prediction challenge
    • BrainPainter (oral at MBIA workshop): a software to generate brain images useful for neuroimaging studies, by e.g. highlighting specific regions according to user input.
    • Disease Knowledge Transfer (poster at main conference): novel method for transferring knowledge across neurodegenerative diseases
  • August 2019: BrainPainter now available straight from the browser:
  • 9-21 June 2019: I will be attending the CVPR and ICML conferences in Long Beach, California. If anyone is around, get in touch!
  • June 2019: Tadpole Challenge Final Results will be announced on Friday 14th June 14:00 BST via YouTube:
  • May 2019: Paper describing the brain visualisation software (BrainPainter) uploaded on arXiv.
  • April 2019: Our paper on the progression of Posterior Cortical Atrophy was accepted in Brain!
  • March 2019: Our paper describing DIVE, the spatio-temporal disease progression model, was published in NeuroImage.
  • February 2019: Started postdoc at MIT with Polina Golland working on neroimaging analysis of stroke.
  • January 2019: Defended my PhD thesis at UCL. Stanley Durrleman and Janaina Mourao-Miranda were part of the examining committee.

Recent Publications


Project’s we’re currently focusing on: Molecular Dynamics, Differentiable simulators, ML Compositionality and Generative Modelling. For prospective students, look at these in particular.

Simulating a Virtual Cell

Simulating a virtual cell using ML-derived coarse-grained potentials

ML Compositionality

ML Compositionality refers to the idea of building a large ML model from modular and reusable building blocks, just like LEGO.

Image Reconstruction

Using ML and compressed sensing techniques to improve the quality, speed and cost of medical scans.

Generative Modelling

Develop models for generation, reconstruction and manipulation of images, text or other high-dimensional data.

ML Benchmarks

We build benchmarks and organize community challenges on key medical prediction problems.

Disease progression modelling

Modelling the progression of Alzheimer’s disease and related neurodegenerative diseases

Differentiable Simulators

Building MRI/PET/Diffusion/MD simulators in PyTorch that can enable us to perform backpropagation through the entire simulator

ML for Molecular Dynamics

Using AI/ML for scaling Molecular Dynamics simulations of proteins.

Medical Visualisation

Building software and ML models for the visualisation of medical images. An example project is BrainPainter.

Lab members

PhD students

Undergraduate students

  • Kevin Bachelor
  • Anderson Compalas
  • Arshia Kapil
  • Anusha Pai
  • Daksh Shah
  • Arthur Wei
  • Sanya Murdeshwar
  • Justin Bui
  • Rahul Nadkarni
  • Jonathan Vengosh
  • Ariel Raizman
  • Jane Choi
  • Alex Feghhi


Sreevani Suvarna: Software Engineer at ADP
Jueqi Wang: PhD student at BU
Junya Ihira: finished his exchange program at UCSC in 2023, returned to Japan
Bhrigu Garg: graduated in 2023

Recent & Upcoming Talks


Lecture Recordings

CSE140 Intro to AI, Winter 2023 (17 lectures):

CSE242 Machine Learning, Fall 2022 (18 lectures):


Romanian TVR (Oct 2023)

Join My Lab

  • Postdocs: Email me directly. I hire all year round.
  • PhD students: Submit a PhD application through UCSC grad amissions. The application deadline is in early January of every year. You are welcome to email me earlier to enquire about our lab or to visit us.
  • UCSC undergraduates/Masters students: I can supervise you on a thesis project if we come up with a very interesting idea, or if you propose your own project. You’re welcome to email me or my PhD students and ask them to suggest a project, or to brainstorm ideas. Another option is to start by helping one of my exiting PhD students on their project. Also see the projects and their associated readings.
  • Visiting students: If you’re a PhD student or postdoc from another research lab and want to visit us for 6-months or 1-year to gain more skills or collaborate, send me an email.