Current stage
Quatm already provides standardised, fully automated experiment operation. The experimentalist writes the experiment description; the framework handles:
- instrument initialisation and state checks;
- multi-dimensional parameter scans;
- real-time acquisition and analysis;
- storage and visualisation;
- queueing and unattended operation.
Next: AI-assisted quantum laboratories
We are training domain models for the daily work of a quantum-physics laboratory. The goal is a system that can understand and execute:
| Capability | Objective |
|---|---|
| Instrument operation | control lasers, magnets, cameras and RF sources autonomously |
| Parameter optimisation | tune experiment parameters from live feedback |
| Fault diagnosis | detect abnormal states and execute recovery procedures |
| Experiment design | generate sequences from a physical objective |
| Data analysis | select and configure analysis pipelines |
This work builds on Quatm’s structured experiment descriptions, device model and data archive: the AI layer operates on well-defined interfaces rather than ad-hoc scripts.
Milestones
| Year | Status | Milestone |
|---|---|---|
| 2024 | Complete | First stable release; open-source core |
| 2025 | Complete | Standardised, fully automated experiment operation |
| 2026 | Complete | Web remote-control platform: browser console, mobile support and bidirectional synchronisation; AI model training begins |
| 2027 | In progress | Validate and iterate the AI model in simulation and controlled experiments |
| 2028 | Target | Quatm AI preview: autonomous, unattended laboratory operation |
Join the roadmap
- Propose directions in GitHub Issues.
- Follow release notes for progress.
- Contact sales@quatm.com for commercial collaboration.