The problem
Quantum computing education is abstract and heavily theoretical. Hardware access is scarce and expensive, simulators return raw numbers without diagnostic insight, and unconstrained AI can hallucinate non-unitary states or unverified mathematics.
Our idea
QubitLearn AI combines visual drag-and-drop circuit design, real-time multi-backend simulation, non-destructive Hilbert statevector tracing, optical circuit recognition, and four-tier trust classification in one interactive learning environment.
What is included
- Interactive quantum studio with a six-basis state cycler and step inspector.
- Statevectors, density matrices, purity, and Bloch sphere projections at every gate slice.
- Hand-drawn circuit photo-to-code scanning with bounding-box error detection.
- Lessons for Deutsch–Jozsa, Grover, Simon's, QFT, VQE, and QAOA.
- Dual-agent coding assessment with hidden invariants and fidelity checks.
Innovation and trust
Type A deterministic computation, Type B multi-simulator consensus, Type C grounded citations, and Type D disclosed AI guidance keep generated explanations separate from physical truth. Formal equivalence and Lean 4 checks support proof-based claims.