Ali SaraerToosi
PhD Candidate · Computational Imaging & Scientific ML · University of Toronto
I'm a first‑year PhD student at University of Toronto, working in the Toronto Computational Imaging Group, advised by Prof. Aviad Levis. I build interpretable machine‑learning models for science. My work spans black‑hole imaging with the Event Horizon Telescope, astrophysics, neural operators, and efficient inference for partial differential equations.
Research
Fuses neural implicit fields with DMD to reconstruct hours‑long EHT movies from <1 % Fourier data.
First invertible RIM that descatters refractive turbulence while preserving intrinsic EHT morphology.
First benchmark for semi‑analytical RIAF models against GRMHD simulations; sets the general approach for further misspecification‑aware uncertainty quantification in black hole imaging.
Proposes a general approach to extend classical KAM theorem to Quantum systems.
Introduces ALINet, a VAE‑based generator that accelerates RIAF parameter inference 10,000-100,000 × over classical sampling.
Industry Experience
Research Scientist Intern1QB Information Technologies, Inc.2023-2024
Recurrent-NN decoder for surface-code error correction in near-zero temperatures.