In a constantly accelerating technological landscape, Lekton.AI is broadening its strategic horizon by focusing on one of the most promising frontiers of innovation: Quantum Machine Learning (QML). This choice stems from a conscious, adaptive reflection on the transformations underway in the ecosystem of advanced research and development, with the aim of strengthening the company’s competitive position in areas of very high scientific and industrial value.

 Hybrid architectures and new development directions

The path Lekton.AI has taken in QML is based on the desire to explore hybrid quantum neural architectures, both convolutional (Quantum CNN) and generative (Quantum GAN), as well as the design of new algorithms aimed at improving the computational efficiency of quantum circuits. Among the main directions is also the investigation of the horizontal scalability of quantum resources through the coupling of heterogeneous systems, a crucial theme for making quantum technologies applicable to real and complex scenarios.

Why Quantum Machine Learning

The integration of QML into Lekton.AI’s strategic lines was not foreseen in the initial plan. However, the growing maturity of the field and the opportunities offered by computationally intensive applications — such as optimisation in financial markets and modelling in the technological-industrial domain — have made this choice not only opportune but necessary. The aim is to differentiate the company’s positioning by focusing on market segments characterised by strong demand for radical innovation.

Academic ecosystem and technology transfer

The new direction has been supported by the collaboration with the University of Siena, and in particular with the PhD programme in Physics, which represents a strategic resource for Lekton.AI. The involvement of PhD students and researchers with expertise in quantum computing, machine learning and theoretical physics allows projects to be grounded on solid scientific foundations, paving the way for the development of genuinely innovative and transferable solutions.

The reference platforms: Qiskit and PennyLane

To support its research activities, Lekton.AI has selected the Qiskit (IBM) and PennyLane (Xanadu) platforms as its main development environments. Both offer advanced tools for hybrid classical-quantum programming, integration with real hardware and simulators, and the design of custom models for quantum neural networks, parametrised circuits and large-scale optimisation techniques.

Towards a new positioning

With this new line of development, Lekton.AI addresses an audience of research institutions, financial institutions and large technology companies — organisations with the structure and resources needed to adopt solutions based on Quantum Computing. In this context, Quantum Machine Learning takes shape as an enabling technology destined to transform entire computational paradigms.

---

Code and repositories

Our QML research does not stay on paper. Part of the work is public and reproducible on our GitHub:

• qh_algorithms — a rigorous experimental comparison between a classical CNN, a hybrid quantum-classical CNN (HQCNN) and a purely quantum reference. The campaign uses multi-seed replicas (R = 10), Wilson confidence intervals, paired Wilcoxon tests and bootstrap to give an honest measure of the variability across architectures. Built with Qiskit and PennyLane on the EuroSAT dataset, it includes a noise-resilience study on 4 qubits. MIT licence.
• QC-2025-26 — the notebooks of our Quantum Computing lab: from the circuit library and the Bloch sphere representation through to the Deutsch-Jozsa, Grover, Quantum Phase Estimation and Shor algorithms, running on the Aer simulator and on real hardware.

A first concrete application of this approach — the comparison between classical and quantum AI — is described in our article on AI in the service of the early diagnosis of Parkinson’s disease.

These repositories are part of the research work linked to the doctoral thesis “New Perspectives on Quantum Technologies: Progress on Quantum Sensing and Quantum Computation” (University of Siena).