An apprenticeship for hardware-aware quantum circuit design
The Qiskit ecosystem teaches Qiskit. That is a different thing from teaching you how a quantum computer actually behaves. Qiskit tutorials show you QuantumCircuit, transpile, Sampler, the cloud submission. They do not show you what happens when your seven-qubit circuit becomes a forty-three-gate transpiled monster because the target backend has the wrong native gate set and your CNOT pattern routes through three SWAP cycles. They do not show you what the same circuit looks like under realistic noise. They especially do not show you how to design a circuit in the first place such that this doesn't happen.
QuantumLearning is the resource for that gap.
The original notebook set covered Qiskit bootcamp material — clean tutorial-level circuits, fine for a weekend. That was never the target. The actual target is backward-designed from "independent hardware-aware quantum circuit designer" and works backwards through what an apprenticeship for that role needs to look like. The teaching spine alternates deliberately between two modes:
- Ideal mode — clean circuit construction, statevectors, probability distributions, the conceptual scaffolding without anything in the way. This is where ideas live.
- Reality mode — the same circuit run through
transpile()with realistic coupling maps, the gate-count blow-up made visible, noisy simulation that shows what an actual IBM-class backend would have produced. This is where designs get tested.
You alternate constantly. Build in ideal, validate in reality. Build, validate. Build, validate. The point isn't the circuits — the point is teaching the reflex of "I just wrote down something beautiful in 6 qubits; what does it cost on a real device?"
The notebooks are now intentionally literature-heavy. Long-form lecture writing inside the notebooks, not only in side documents. Multiple-choice concept checks in every technical notebook. Editable circuit labs with widget-based parameter sweeps. None of it requires an IBM Quantum account, a cloud token, or a live API dependency. Aer simulator, locally, fully reproducible.
I keep finding that the only reliable thing about learning quantum computing is that the difference between "I understand this concept" and "I can design a working circuit that does this concept on a real device" is enormous and almost never addressed by curriculum. People emerge from textbooks knowing Bell states and emerge from cloud tutorials knowing how to call Sampler — neither lets them sit down and design something new.
This material aims at that gap.
What becomes possible: an amateur ends the course able to read a quantum hardware roadmap and know which proposed algorithms map onto which generations of hardware.