Research
Distributed quantum computing
Running one quantum circuit across several processors joined by quantum links.
A single quantum processor will not hold the qubits that large applications need, so several processors are linked over quantum links and run as one machine. Executing a circuit on such a system needs a compiler that maps logical qubits to physical ones across processors and then routes them, generating entanglement, scheduling teleportations and inserting swaps, under link fidelity, decoherence and congestion constraints.
Compilers
From a circuit to physical operations across several processors.
A quantum circuit is written without reference to the hardware that will run it, and compilation closes that gap in two stages. Gate-level compilation rewrites the circuit in the native gate set of the machine. Hardware-level compilation then turns those gates into operations on the physical device. The two can be optimized jointly, but this work focuses on the hardware level, where the task is twofold. Mapping assigns logical qubits to physical qubits across the processors, anticipating which gates will have to act across a link. Routing then brings qubits together where the gates need them, generating entanglement between processors, scheduling gate and qubit teleportations and inserting swaps to move qubits within and across processors, all under link fidelity, decoherence and congestion constraints.
With Cisco Systems I built compilers for both gate-based and measurement-based distributed execution, including a reinforcement learning compiler that minimizes expected execution time and a genetic approach that minimizes the number of gate and qubit teleportations. The demo below showcases the modeling environment proposed in the ICC 2025 paper, on a two-processor system.
Papers