Campus Directory

Division Baskin School of Engineering

Department/ College/ Unit Computer Science and Engineering

Title Assistant Professor

Affiliation Faculty

Faculty Type Regular Faculty

Areas of Expertise Machine Learning
Artificial Intelligence
Mathematical Modeling

Web Site

Primary Office Location Engineering Building 2
547A

Mail Stop No Mailstop

Biography, Education and Training

Razvan Marinescu is an Assistant Professor of Computer Science and Engineering at UC Santa Cruz. His research combines machine learning, medical imaging, and protein simulation to understand disease and advance healthcare. He develops generative models, Bayesian inference methods, and differentiable simulators for medical image reconstruction, disease progression modeling, and AI-assisted drug discovery. His work links the mathematical foundations of AI with applications in biology and medicine.

Before joining UCSC in 2022, he conducted postdoctoral research in Polina Golland’s laboratory at MIT’s Computer Science and Artificial Intelligence Laboratory. He received his Ph.D. in Computer Science from University College London in 2019 and his M.Eng. (2014) in Computer Science from Imperial College London. His doctoral research developed Bayesian models of Alzheimer’s disease and related neurodegenerative conditions.

He co-organized the international TADPOLE Challenge for predicting Alzheimer’s disease progression and served as president of the MIT Postdoctoral Association. His entrepreneurial work includes co-founding GiwoTech, which develops AI for drug discovery, and Lumos AI, which develops evaluation and benchmarking tools for healthcare and life sciences.

Summary of Expertise

Machine learning and artificial intelligence for medicine, biology, and scientific computing. Expertise includes generative modeling, Bayesian statistical inference, inverse problems, medical image reconstruction, and computational modeling of neurodegenerative disease. Additional areas include differentiable physics simulators, machine learning for molecular dynamics and drug discovery, mathematical foundations of diffusion models, and evaluation and interpretability of medical large language models. This work combines mathematical modeling with computational methods to infer biological structure, understand disease processes, and develop more reliable AI systems.

Research Interests

• Foundations of generative AI: Diffusion models, flow matching, Bayesian inference, and connections between machine learning, stochastic analysis, Malliavin calculus, and physics.

• Differentiable simulation and inverse problems: Inferring physical and biological parameters from observations using differentiable simulators, probabilistic models, and numerical optimization.

• Medical imaging and virtual brain models: MRI and CT reconstruction, low-field MRI, brain microstructure inference, and digital twins for studying neurological disease.

• AI for molecular simulation and drug discovery: Coarse-grained molecular dynamics, neural network potentials, enhanced sampling, protein conformational dynamics, and multiscale approaches toward whole-cell simulation.

• Disease progression modeling: Longitudinal and multimodal models of Alzheimer’s disease and other neurodegenerative disorders, including approaches that integrate imaging, clinical measurements, and biological data.

• Medical large language models: Evaluation, benchmarking, interpretability, and analysis of how models represent and use medical knowledge.

Teaching Interests

Undergraduate and graduate teaching in machine learning, artificial intelligence, probabilistic modeling, and scientific computing. Particular interests include the mathematical foundations of learning, generative models, Bayesian inference, optimization, and applications of AI to biology and medicine.

Courses taught at UC Santa Cruz include:

• CSE 140: Artificial Intelligence

• CSE 142: Machine Learning

• CSE 242: Machine Learning — graduate level

• CSE 290C: Advanced Topics in Machine Learning

Awards, Honors and Grants

Selected grants and research support

• NSF STTR Phase I, 2025: “Accelerating Molecular Dynamics using Coarse-Grained Neural Network Potentials,” a GiwoTech–UC Santa Cruz collaboration. Total project award: $305,000; NSF award 2451680.

• NSF Innovation Corps (I-Corps): $50,000 award for customer discovery and market research in the biopharmaceutical industry.

Selected awards and honors

• 2021 — Best Paper Award, NeurIPS Workshop on Deep Generative Models and Downstream Applications, for “Bayesian Image Reconstruction using Deep Generative Models.”

• 2017 — Joint runner-up, François Erbsmann Prize, International Conference on Information Processing in Medical Imaging (IPMI).

• 2011 — Best Undergraduate Project in Artificial Intelligence, Imperial College London.

Advisees, Grad Students & Researchers , (nmashhad)
, (enikolak)
Gupta, Utkarsh (utgupta)
Khole, Prathamesh Pradeep (pkhole)
, (dsabo)
Terry, Ian (imterry)
Mistreanu, Ionut (imistrea)
, (acbruce)
Mirafzali, Ehsan (smirafza)

Courses CSE242
CSE140
CSE290C