At the Quantum Compilation Group of Fraunhofer IIS in Nuremberg we are searching for two students working on the topics described below.
Please apply directly via Working Student (all genders) – Quantum Computing / Quantum Circuit Compilation at Fraunhofer-Institut für Integrierte Schaltungen IIS | softgarden and append your CV and transcript.
Learning‑Based Quantum Compilation for Quantum Chemistry
Building on recent advances in quantum compilation for quantum chemistry, the project reformulates the optimization of physically inspired circuits for hydrogen chains as a learning‑based compilation task. The goal is to minimize circuit depth and two‑qubit gate count while preserving chemical accuracy and fidelity, thereby enabling execution on contemporary quantum hardware. You will design the ML/RL setup (state representation of circuits and molecular targets; permitted transformations such as symmetry/tapering, commutation/cancellation, and topology‑aware routing/scheduling; and a reward capturing accuracy‑vs‑cost trade‑offs), implement efficient simulations, and benchmark standard methods such as PPO, with optional extensions to search‑based techniques like MCTS. Should the results warrant, we aim to develop the work into a manuscript for publication.
This topic is available as a master’s thesis.
Requirements: proven experience in both the theory and implementation of machine learning and reinforcement learning; familiarity with quantum computing and libraries such as Qiskit/Qiskit Nature or PennyLane, and with quantum‑chemistry workflows; prior exposure to circuit optimization/transpilation and hydrogen‑chain benchmarks is a strong plus.
Reinforcement Learning for Tailoring Quantum Error Correction to Structured Noise
Building on Meyer et al. [https://doi.org/10.48550/arXiv.2506.11552], the project reformulates the search for quantum error correction encodings—originally approached with variational routines and simple gradient methods—as a reinforcement learning problem. The goal is to learn codes that minimize information loss under structured noise, and thereby improve resource-efficiency of quantum error correction. You will design the RL setup (MDP consisting of state representation, available operations, and reward signal) for efficient simulation, and benchmark standard methods such as PPO, with optional extensions to advanced methods like MCTS. Should the results warrant, we aim to develop the work into a manuscript for publication.
This topic is available as either a master’s thesis or a student assistant position.
Requirements: proven experience in both the theory and implementation of reinforcement learning; familiarity with quantum computing and libraries such as Qiskit or PennyLane; prior exposure to quantum error correction is a strong plus.