Many thanks to my collaborators Sergio Barrera Cabodevila and Aleksi Kurkela!

In this paper, we use machine learning tools to train a neural network to evaluate the collision terms in QCD kinetic theory simulations, surpassing the speed of traditional methods. This could pave the way for event-by-event modeling of heavy-ion collisions using QCD kinetic theory.

Thanks a lot to my collaborators João Barata, Kirill Boguslavski and Andrey V. Sadofyev!
Check out the paper at arxiv.org/abs/2511.07519

We find that the nonequilibrium early stages imprint a signature on the substructure within the jet. These results provide a foundation for more accurate descriptions of hard-probe evolution in heavy-ion collisions, opening new avenues for understanding QCD matter under extreme conditions.

In this paper, we computed how energetic partons (jets) radiate soft gluons while propagating through dynamically evolving QCD matter using the Improved Opacity Expansion, taking into account the medium evolution during the emission process.

When heavy ions collide at large energies, they create a rapidly evolving, out-of-equilibrium plasma of quarks and gluons, the quark-gluon plasma. Most jet quenching models assume that this medium is static or already thermalized, missing the dynamics during the early, far-from-equilibrium stages.

Many thanks to my collaborators Kirill Boguslavski, @aleksasmaz.bsky.social, Adam Takacs and Fabian Zhou!

Check out the paper at arxiv.org/pdf/2510.25669

Using QCD kinetic theory simulations, we analyze and compare the minijet evolution with jet transport coefficients. We find that the thermalization time scales remarkably well with these transport coefficients, enabling a phenomenological estimate of the minijet quenching time.

Heavy-ion collisions produce a hot, dense state of matter known as the quark-gluon plasma. In this paper, we study how energetic particles (‘Minijets’) interact with this medium, how they lose energy and eventually thermalize.

opening up new questions about how such processes should be modeled in kinetic theory simulations, and highlighting interesting directions for further study.

I'm very grateful to my collaborators Alois Altenburger and Kirill Boguslavski!

Check out the paper at arxiv.org/pdf/2509.03868

the phenomenon where jets lose energy while traveling through the medium.
As an application, we also computed the rate of a gluon in the plasma to emit another gluon. This differs from commonly used approximations,

In this paper, we use QCD kinetic theory to calculate the elastic collision kernel, which describes how likely it is for a jet particle to receive momentum kicks from the plasma. This kernel is an essential ingredient for understanding jet quenching,

When two heavy ions collide at very high energies, they create a hot, dense plasma of quarks and gluons that is initially far from equilibrium. We studied how this early plasma interacts with energetic particles (“jets”) that pass through it.

traditional methods. This could pave the way for event-by-event modeling of heavy-ion collisions using QCD kinetic theory.

Many thanks to my collaborators Sergio Barrera Cabodevila and @aleksikurkela.bsky.social

Check out the paper at arxiv.org/abs/2506.19632

However, these simulations are computationally expensive due to the high-dimensional collision integral that appears in the effective Boltzmann equation. In this paper, we use machine learning tools to train a neural network to obtain the collision integral much faster than