Quantum Transportation, a majority-owned subsidiary of Rail Vision, delivered a working integration layer that ingests Google's public experimental surface-code dataset into its patent-pending quantum error correction (QECC) transformer pipeline. The integration uses a standardized data adapter and dynamic attention masking to accommodate varying code distances and layouts.

The team implemented an end-to-end training loop capable of processing mixed batches of real experimental shots, building on the subsidiary’s previous AWS cloud deployment of its transformer-based neural decoder. This pipeline ingests both internal and external data formats, supporting continuous training and validation across diverse test configurations.

By enabling repeatable benchmarking on a credible external testbed, the milestone reduces technical risk and lays the groundwork for scalable quantum error correction applications. The enhanced pipeline positions the company to commercialize advanced neural decoders for industries such as transportation, computing, and data processing.