Quantum software developer Kipu Quantum has made its Rimay quantum feature extraction service available in the IBM Quantum Functions Catalog, giving IBM Quantum Flex, Premium, and On-Prem Plan users direct API access to quantum-enhanced machine learning within existing data science pipelines. Following the earlier catalog release of Kipu’s Iskay Quantum Optimizer, Rimay expands the company’s commercial presence on IBM Quantum hardware from combinatorial optimization into tabular and multimodal feature engineering.
| Kipu Quantum Rimay Function Specifications & Hardware Benchmarks | ||
|---|---|---|
| System Dimension | Operational Specification | Measured Benchmark / Platform Impact |
| Algorithmic Method | Digitized counterdiabatic evolution of spin-glass Hamiltonians | Extracts single-body expectations and two-body qubit correlations for classical classifiers. |
| Workload Capacity | Up to 156 features (1 qubit per feature) & 15,000 samples per job | Supports Small (≤2,999), Medium (3,000–9,999), and Large (10,000–15,000) compute tiers up to 10M shots. |
| Nighthawk r2 Runtime | 48-qubit German credit dataset (1,000 samples, 5,000 shots) | Cuts QPU usage from 22 min on Heron r3 (ibm_boston) to 2 min on Nighthawk r2 (ibm_phoenix). |
| Classifier Uplift | Hybrid Random Forest (classical + quantum features) | Increases recall from 0.921 to 0.957 and AUC from 0.781 to 0.798 over classical baselines. |
Rather than replacing established machine learning models, Rimay operates as a drop-in preprocessing layer between train-test data splitting and classical model training. During the initial fit-transform phase, the function maps normalized numerical features to individual qubits and encodes pairwise mutual information into Hamiltonian coupling terms. After executing digitized counterdiabatic circuits on the quantum processor, it returns an enriched CSV matrix that teams can feed directly into standard frameworks such as scikit-learn, XGBoost, CatBoost, or LightGBM.
To support repeatable production inference, Rimay stores the trained Hamiltonian topology under a persistent mapping_id bound to the calibrated processor layout. Data science teams can subsequently invoke transform mode on new, unlabeled datasets without re-fitting the feature map, or pair the service with Kipu’s offline classical surrogate framework to eliminate quantum hardware dependencies during high-volume inference.
Kipu Quantum noted that Rimay was recently deployed with a Fortune 100 enterprise on an IBM Quantum Nighthawk processor to enhance customer-intent prediction models on tabular datasets. The release is accompanied by an end-to-end credit-risk classification tutorial in IBM Quantum documentation.
Official announcement details are available via the Kipu Quantum newsroom here and the Rimay product overview here. Catalog access and technical documentation are accessible on the IBM Quantum Functions Catalog here and IBM Quantum Docs here. Supporting scientific preprints on arXiv detail digitized counterdiabatic quantum feature extraction here and quantum-enhanced satellite image classification here. Related QCR coverage includes Kipu Quantum’s initial Rimay launch here, its offline surrogate inference framework here, and IBM’s QDC 2026 Nighthawk roadmap here.
November 13, 2026
In GQI Portal
The players behind the news
The team that writes QCR tracks every company, deal and technology in the GQI Factory, GQI's verified database of the quantum industry. Next up: every story linked to its players, coming to QCR's paid plans.
- Players Companies and institutions across the quantum industry, by segment.
- Scorecards How the players compare on hardware, software, funding and more.
Newsletter
QCR Alerts in your inbox
The latest reporting and analysis from Quantum Computing Report. Free, unsubscribe any time.