Simone Silvestri
Politecnico di Torino. Computational fluid dynamicist and ocean modeler.
DIATI, Politecnico di Torino
Torino, Italy
I am a computational fluid dynamicist with a Marie Sklodowska-Curie Actions (MSCA) fellowship at Politecnico di Torino and a Research Affiliate at MIT, where I keep collaborating with the Climate Modeling Alliance (CliMA). I spend most of my time convincing GPUs to simulate the ocean — and occasionally the ocean cooperates.
I am a core developer within the CliMA organization, where I work on Oceananigans.jl and ClimaSeaIce.jl, and I am a co-owner of the NumericalEarth organization, which aims to build a community around GPU-based tools for Earth system simulation in Julia. I built the dynamical core and numerical schemes that allow to run mesoscale-resolving climate simulations on relatively modest hardware, which is a polite way of saying I optimized things until the electricity bill became the bottleneck.
Before the ocean, there was fire. I obtained my PhD from Technische Universiteit Delft in 2021, working on turbulence-radiation interactions in high-temperature flows and developing a fast GPU Monte Carlo method for radiative heat transfer coupled with DNS. Turns out, the leap from simulating furnaces to simulating oceans is shorter than you would think — both involve turbulence, GPUs, and a healthy disregard for easy problems.
news
| Sep 15, 2025 | Started my MSCA Postdoctoral Fellowship at Politecnico di Torino (DIATI), working on GPU-accelerated ocean modeling and high-order numerical methods for geophysical turbulence. |
|---|---|
| May 1, 2025 | Our GPU ocean dynamical core paper was featured as an EOS Editor’s Highlight: A Leap Toward Next-Generation Ocean Models. |
| Jul 1, 2024 | Our paper on a new WENO-based momentum advection scheme for ocean mesoscale turbulence was published in JAMES. |
selected publications
- A GPU-Based Ocean Dynamical Core for Routine Mesoscale-Resolving Climate Simulations (2025)
- A New WENO-Based Momentum Advection Scheme for Simulations of Ocean Mesoscale Turbulence (2024)
- Oceananigans.jl: A model that achieves breakthrough resolution, memory and energy efficiency in global ocean simulations (2023)
- Turbulence modulation in thermally expanding and contracting flows (2021)
- A fast GPU Monte Carlo radiative heat transfer implementation for coupling with direct numerical simulation (2019)