Research
My research spans machine learning for gravitational-wave astronomy and foundational problems in classical general relativity.
My group’s main focus is developing simulation-based inference methods for gravitational-wave data analysis, enabling fast-and-accurate parameter estimation with neural networks. In classical gravity, I have worked on several distinct problems: black hole perturbation theory (where recent results on quasinormal mode orthogonality connect to ringdown modelling) as well as the nonlinear dynamics of anti-de Sitter spacetime and the cosmological backreaction problem.
See also the research group and publications.
Machine Learning for Gravitational Waves
Gravitational-wave detectors are observing an increasing number of signals from merging black holes and neutron stars, and extracting the physics from each event requires computationally intensive Bayesian sampling techniques. We develop simulation-based inference methods that train neural networks to learn the posterior distribution from simulated data, bypassing the likelihood entirely.
Our open-source code Dingo produces full posterior distributions in seconds (compared to hours or days for traditional samplers) and has been applied for accurate inference of many LIGO-Virgo-KAGRA events, including binary neutron star mergers (Nature, 2025). We have also recently extended to population-level inference.
Classical Gravity
It’s essential to have accurate theoretical predictions for interpreting gravitational wave data, especially as observations become more precise with sensitive detectors. My main focus is on the ringdown problem—the characteristic oscillations following a binary merger.
Recent highlights include uncovering an orthogonality relation between quasinormal modes (putting them on a closer footing to normal modes) and developing metric reconstruction for sourced Teukolsky solutions in Kerr. Applications include to boson clouds, nonlinear ringdown, and gravitational turbulence. In past work, I studied gravitational turbulence in anti-de Sitter spacetime and established a rigorous backreaction framework ruling out large-scale effects from cosmological inhomogeneities.