WestQuant Open is an open-source initiative focused on AI-guided optimization of quantum algorithms and mathematical representations.
The goal is to build an open layer that can work across existing quantum ecosystems such as Qiskit, CUDA-Q, PennyLane, TKET and Cirq. Rather than replacing these platforms, WestQuant explores how AI can help decide which mathematical representation, transformation and optimization path to use before and during compilation.
A central part of the project is WQT20, a compact 20M-parameter Transformer being developed specifically for quantum representation scheduling. WQT20 is designed to learn from mathematical structures, quantum circuits, Hamiltonians, hardware constraints and optimization trajectories, while deterministic quantum tools execute and verify the transformations it proposes.
The long-term research loop is:
Open Source → Integrations → Users → Experiments → Representation Graphs → Better AI → Better Optimization → Better Benchmarks
Models, research code, benchmarks and integrations will progressively be released through WestQuantOpen on GitHub and WestQuantStudio on Hugging Face.
The ambition is simple: make WestQuant an open research layer that helps researchers and developers explore the representation space of quantum algorithms more efficiently.
WestQuant Algorithm Studio
For our own research we are currently developing WestQuant Algorithm Studio, an AI-assisted research environment for quantum algorithm design, simulation and optimization. The studio already supports an AI-in-the-loop workflow, where our own in-house mathematical AI, WestMath, assists with representation selection, circuit construction, transformation proposals and optimization suggestions during experimentation. This gives us an active software foundation for exploring alternative algorithmic formulations rather than starting from a purely conceptual research idea. Hopefully, we can launch a public version of the Studio somewhere down the road.
WestMath is an earlier version of WestQuantTransformer20M (WQT20), larger in size and wider in applications – Like the mother and a mathematical foundation model. The WQT20 we are launching as open source is a compact, easy to use and powerful mathematical AI model optimized for Quantum Representation Scheduling and Quantum Algorithm Optimization. The essence of several years of mathematical development of AI models. H
