Welcome to my homepage!
I am currently a Gordon and Betty Moore postdoctoral research fellow at the Physics Department, University of California (Berkeley). I study the dynamics of many-body quantum and classical systems away from equilibrium. My research focusses on Condensed Matter Physics, Ultracold Atoms, Statistical Mechanics, and the interplay between Machine Learning and Physics.
You can learn more about me and my research on this website.
Open PhD and Postdoctoral Positions
PhD and postdoc fellowships at the interface of quantum mechanics and machine learning at Sofia University starting October 1, 2020. Apply now!
Geometric Speed Limit of Accessible Many-Body State Preparation
A physical way to characterize the complexity of preparing quantum states is the quantum speed limit – the minimal time required to prepare a target state with unit fidelity. We formulated a Conjecture according to which, for any protocol within the space of available controls, the time-integrated energy fluctuations are bounded from below by the adiabatic geodesic length between the initial and final states.
Reinforcement Learning to Control the Kapitza Oscillator
A Reinforcement Learning agent figures out how to Control Floquet-Engineered Quantum States in a numerical simulation of a quantum experiment. The ability of RL to control systems far away from equilibrium is demonstrated by steering the quantum Kapitza oscillator into the stabilized inverted position in the presence of a strong periodic drive.
Parametric Instabilities in Periodically Driven Bosonic Systems
Floquet engineering is an established technique from the quantum simulation toolbox, with the present bottleneck set by unwanted heating. We study the occurrence of Parametric Instabilities in Weakly-Interacting Bosonic Optical Lattices, and showcase the detrimental effects caused by transverse degrees of freedom and weak harmonic traps.
Glassy Phase of Quantum Control
We present evidence for a universal spin-glass-like transition in the quantum control landscape, controlled by the control protocol duration. The glassy critical point is marked by a proliferation of protocols with close-to-optimal fidelity and with an optimal protocol that appears exponentially difficult to find.
A High-Bias, Low-Variance Introduction to Machine Learning for Physicists
A comprehensive Review Paper on Deep Learning techniques and concepts useful in physics: “A High-Bias, Low-Variance Introduction to Machine Learning for Physicists”.
Reinforcement Learning Quantum State Preparation
A Reinforcement Learning agent learns to Prepare Quantum States in a non-integrable Ising chain. With no prior knowledge about the quantum system, the agent learns to extract the essential features of the optimal preparation protocol.
Phase Transitions of Quantum Control
Control problems are optimisation tasks with the objective to find the most efficient solution satisfying given constraints. We demonstrated that the difficulty of finding the optimal solution can change abruptly, resulting in the appearance of Control Phase Transitions in the optimisation landscape.
Periodically Driven Systems
Recent surge of activity in Periodically Driven many-body Systems has spawned a new branch of quantum physics, known as Floquet engineering: the synthetic generation of novel Hamiltonians, otherwise inaccessible in static condensed matter systems.
The ability to simulate quantum systems offers invaluable opportunities for building up intuition and understanding about the bizzare world of quantum mechanics. We developed QuSpin – an open source Python package for exact diagonalisation and dynamics of quantum many-body systems.