Simulation plays a vital role in advancing quantum physics, where direct experimentation is often limited by cost, fragility, and technological feasibility. However, traditional simulations frequently rely on idealized assumptions that limit their predictive accuracy in real-world conditions. This project addresses this gap by developing an integrated robotic platform for autonomous quantum experimentation. We use a Franka Research 3 robotic arm, which will interface with a QuantenKoffer photonic toolkit to construct and reconfigure optical quantum experiments. Using reinforcement learning, agentic programming, and physics-informed world models, the system will propose, simulate, and refine experimental designs in a closed feedback loop. GPU-accelerated simulations will generate high-fidelity priors to guide reinforcement learning toward physically realizable outcomes, such as maximizing entanglement fidelity. By coupling robotic automation with simulation-informed learning, this work aims to create a self-improving system that bridges the divide between theoretical predictions and experimental implementation, accelerating the design and execution of quantum experiments.

I am super excited to join Dr. Englund’s lab and explore how digital twins and reinforcement learning can transform quantum optics experimentation. By combining Omniverse-based simulations with physics-informed models and robotic automation, I hope to learn how to bridge the gap between idealized theory and noisy, real-world systems, ultimately accelerating the design of reliable, autonomous quantum experiments.