Intracluster Gas Physics

Translating telescope observations into cosmological constraints requires solving a chain of non-trivial inference problems. Each galaxy cluster is an extended source with a complex 3D gas distribution; we observe it projected on the sky, mixed with noise, filtered by the instrument response, and potentially contaminated by other astrophysical signals. Extracting reliable measurements from this data - and propagating all sources of uncertainty through to final constraints - demands careful forward modeling, Bayesian parameter estimation, and a thorough understanding of systematic effects.

My PhD work was part of the NIKA2 SZ Large Program (LPSZ), a survey of ~40 galaxy clusters with the NIKA2 camera, a state-of-the-art millimeter-wave instrument at the IRAM 30m telescope. I led the end-to-end analysis of individual clusters: from raw time-ordered detector data to calibrated sky maps, and from maps to inferred thermodynamic profiles using MCMC. I also developed a hierarchical Bayesian framework used to combine individual cluster measurements into population-level scaling relations, with full propagation of measurement uncertainties, selection effects, and mass estimation biases.

Technical methods: MCMC (Metropolis-Hastings and gradient-free samplers), forward modeling with instrument transfer functions, joint multi-wavelength modeling (millimeter + X-ray), hierarchical Bayesian regression, Monte Carlo forecasting.