Internship Focus: Machine Learning & Quantum Feedback for Cat Qubits Context: Cat codes are a promising approach for quantum error correction in superconducting quantum processors. As systems scale, optimizing control and measurement becomes complex, especially with parameter drifts (e.g., magnetic flux, drive amplitude). Recent work in the group has shown the potential of reinforcement learning for cat state preparation and autonomous error correction. New Opportunity: The lab now has an Nvidia DGX Quantum instrument, enabling low-latency (µs) between a powerful server and quantum devices. This allows real-time training of large neural networks for quantum feedback previously impossible. Internship Tasks: Optimize CNOT Gates: Use machine learning to find optimal driving sequences for CNOT gates between cat qubits, and identify unknown parasitic terms in the system’s Hamiltonian. This is part of the RobustSuperQ project (France 2030, PEPR Quantique). Quantum Trajectories: Implement a recent theoretical breakthrough: computing the quantum state of a continuously monitored system for finite time intervals (not just infinitesimal). Design experiments to demonstrate this on superconducting circuits, with the goal of real-time trajectory computation and adaptive control using the DGX Quantum. Real-Time Parameter Estimation: Develop and implement real-time parameter estimation and feedback correction for cat qubit devices.