What does an accurate model of the universe consist of? First, it must rely on the correct fundamental laws of physics, as we know them today. Second, because we only observe the universe from a specific vantage point at a specific time in its history, we need to make some assumptions about the symmetries and large-scale
properties of the universe to be able to extrapolate from our observations to what distant regions of the universe might look like. Finally, once the theatre has been set by the fundamental laws and the cosmological hypotheses, we need to describe all the observations we make in models that are detailed enough to not introduce theoretical biases in our interpretation and analysis of observations. This accurate modelling of small but subtle effects is also essential to help uncover new phenomena by pointing to observables challenging minute aspects of the model.
Over the years, I have been involved in investigating all three of these pillars of cosmology with a constant focus on trying to test their limits and robustness. In this defence, I will specifically focus on:
1. tests of the cosmological principle;
2. light propagation in a lumpy universe;
3. line-of-sight effects in strong gravitational lensing.

Ultra-light bosons such as axions and dark photons are copiously produced around black holes due to the superradiant instability, at the expense of the black hole’s mass and spin. Here I present two new signatures of black-hole superradiance that can be used to look for these particles. First, I show that the superradiant spin-down of super massive black holes is uniquely imprinted in measurements of stellar tidal disruption event (TDE) rates. Upcoming TDE rate measurements by the LSST could reveal or set constraints on the existence of axions or dark photons with masses around 10^-19 eV. And second, I show that kinetically-mixed dark-photon superradiant clouds around stellar mass black holes lead to a new type of electromagnetically bright and possibly pulsating astrophysical source. These “new pulsars”, which are realized for dark photons with masses around 10^-13 eV, can be looked for using X-ray and radio telescopes, and by performing GW searches of anomalous pulsars.

We are living exciting times: we are now able to probe the most violent events of the Universe with diverse messengers (cosmic rays, neutrinos, photons and gravitational waves). One challenge to complete the multi-messenger picture resides in the highest energies, as no ultra-high energy neutrinos have been observed yet. This challenge could be undertaken by the GRAND (Giant Radio Array for Neutrino Detection) project, which aims at detecting ultra-high energy particles, with a colossal array of 200'000 antennas over 200'000 km2, split into ~20 sub-arrays of ~10'000 km2 deployed worldwide. In this talk, we will present preliminary designs and simulation results, plans for the ongoing, staged approach to construction, and the rich research program made possible by the proposed sensitivity and angular resolution.

After giving a short introduction on gamma-ray bursts (grb) prompt phase, I will discuss two aspects of the presence of protons in the jet. Firstly, I will present a model of emission based on proton synchrotron and the production of pairs via the Bethe-Heitler process. Secondly, after reviewing why grbs are good candidates to be the site of ultra-high energy cosmic ray (UHECR, around 10^20 eV) acceleration, I will show how the characteristics of grb emission constrain the maximum energy that UHECR can reach in their jet to be less than 10^19 eV for fiducial grb parameters. Finally, I will explain why a specific class of grbs, namely low-luminosity grbs, could still satisfy the constraints. However, I will show that their contribution to the total flux of UHECR is necessarily limited

A dark energy-like component in the early universe, known as early dark energy (EDE), is a proposed solution to the Hubble tension. While in an MCMC analysis, the EDE model is disfavoured by Cosmic Microwave Background and large-scale structure data, a frequentist profile likelihood analysis prefers consistently larger amounts of EDE and with that a Hubble constant consistent with low-redshift probes. The difference between MCMC and profile likelihood can be explained by prior volume effects in the MCMC analysis. I will describe how a frequentist profile likelihood yields important complementary information compared to a Bayesian MCMC analysis.

Gamma-ray bursts (GRBs) are one of the most energetic explosions known in the Universe and are also known to have the most relativistic jets, with initial expansion Lorentz factors of hundreds. Depending on the prompt duration and hardness ratio, GRBs are broadly classified as hard-short (<2 s) and soft-long (>2 s) GRBs. The prompt phase is followed by the afterglow, observed for several days at all wavelengths, from radio to TeV including optical and X-ray bands. The origin of the plateau phase observed in the early X-ray light curve of GRBs has been debated since its discovery in 2005. In our recent paper, we show that observed plateau phase can be explained within the classical GRB model by considering an expanding stellar shell with initial jet Lorentz factors of tens. In this talk, after giving a brief introduction about the topic, I will present the key findings of the paper and I will list all the supporting evidences for this result.

Cool evolved stars, namely asymptotic giant branch stars and red supergiant stars, experience an important mass loss
(10^-7 to 10^-4 solar mass per year, while the solar wind represents ~10^-14 solar mass per year).
This stellar wind is enriched in heavy elements and contributes to the chemical evolution of the Galaxy.
For massive stars, it can determine the final fate of the star as a neutron star or black hole.
Yet, some of its mechanisms are still unknown: (1) the dust nucleation process is still unclear,
(2) for low and intermediate mass stars the processes shaping planetary nebulae are not well constrained,
(3) for massive stars we still do not know how the material is lifted from the photosphere.
I will present recent high angular resolution results on cool evolved stars using adaptive optics, and optical
and millimeter interferometry.
I will conclude the presentation with an overview of a promising future thanks to a powerful new instrumentation.