About Me

I am a PhD researcher in Quantum Machine Learning (QML) at the University of Basel, working within the Optimization of Machine Learning Systems Group. My research is supported by the SNSF and conducted in collaboration with IBM Research Zürich.

Driven by a passion for mathematics, physics, and computer science, my goal is to leverage abstract mathematical frameworks to solve complex, real-world problems. My interest in computation originated from an early curiosity about how software simulates reality, from digital physics in movies to video game mechanics and design.

Outside of research, I produce music across various genres and styles. You can explore some of my creations on YouTube.

Research

My research lies at the intersection of mathematics, theoretical computer science, and physics, focusing on the rigorous mathematical foundations of Quantum Machine Learning (QML). I leverage theoretical tools from high-dimensional probability, mathematical optimization, statistical learning theory, and computational complexity to investigate three main pillars:

Optimization Landscapes and Trainability: A fundamental barrier in training variational quantum algorithms is the barren plateau phenomenon, where cost function gradients vanish exponentially with system size. While mitigation strategies can improve gradient scaling, they frequently force quantum models into regimes that are classically simulable. My research includes mathematically characterizing this trainability–simulability trade-off.
Boundaries of Classical Simulation: Beyond developing algorithms for simulating quantum circuits, I also explore the exact computational boundaries separating classical simulability from non-classical quantum hardness. A key theoretical conjecture guiding this work is whether eliminating barren plateaus fundamentally implies classical simulability, thereby constraining the scope of provable "quantum advantage" in learning tasks.
Risk Analysis and Generalization: I analyze the generalization properties of quantum learning models, with an emphasis on non-monotonic risk behaviors such as the double descent phenomenon at the interpolation threshold. My objective is to derive exact analytical expressions showing how generalization error depends on data geometry (e.g., covariance eigenspectra) and physical quantum circuit architectures.

Publications

  1. Sabri Meyer, Francesco Scala, Francesco Tacchino, and Aurélien Lucchi. Gradient scalability and Taylor surrogation of quantum cost landscapes. Phys. Rev. Research 8, 023325, 2026.

Preprints

  1. Sabri Meyer, Francesco Scala, Francesco Tacchino, and Aurélien Lucchi. Trainability of Quantum Models Beyond Known Classical Simulability. arXiv preprint arXiv:2507.06344, 2025.

Conferences & Workshops

QIP 2026 — 29th Annual Quantum Information Processing Conference
ATTA Centre, Riga, Latvia
QTML 2025 — 9th International Conference on Quantum Techniques in Machine Learning
Marina Bay Sands, Singapore
LMS Research School 2025 — Quantum Machine Learning and Hamiltonian Simulation
Sabhal Mòr Ostaig, Scotland
QTML 2024 — 8th International Conference on Quantum Techniques in Machine Learning
University of Melbourne, Australia

Teaching

Fall Semester 2023 – Present: Teaching Assistant in Mathematics of Data Science

  • Introduction to probability theory and mathematical statistics
  • Random matrix theory, concentration inequalities, functional calculus, and high-dimensional analysis
  • Marchenko–Pastur law, neural networks, distance distributions, and stochastic processes
  • Graph theory, spectral clustering, and diffusion maps

Curriculum Vitae

2023 – Present
PhD in Quantum Machine Learning
University of Basel
  • Electives in Algebraic Geometry
2021 – 2023
Master of Science in Mathematics
University of Basel
  • Specialized in Algebraic Number Theory, Analysis of PDEs, and Stochastic Analysis
  • Master Thesis on the Hessian Eigenspectrum of Nonlinear Neural Networks
  • Electives in Theoretical Quantum Mechanics
2018 – 2021
Bachelor of Science in Mathematics
University of Basel
  • Electives in Physics & Computer Science