Quantum Developer Foundations.
A practical starting point for developers. Learn the concepts below, use the interactive labs, then demonstrate your understanding in an independent assessment and circuit project.
Represent a quantum state.
A qubit has a state α|0⟩ + β|1⟩, where α and β are complex amplitudes and |α|² + |β|² = 1. Measuring in the computational basis gives 0 with probability |α|² and 1 with probability |β|². An amplitude magnitude of 0.5 therefore contributes probability 0.25, not 0.5.
The |+⟩ state is (|0⟩ + |1⟩)/√2. Both amplitudes have magnitude 1/√2, so each measurement outcome has probability 1/2. The |−⟩ state changes the relative sign. It has the same computational-basis probabilities, but it behaves differently under later gates. A quantum state carries information that a probability histogram alone does not show.
Two qubits need four amplitudes. An exact n-qubit statevector needs 2ⁿ complex entries, so one additional qubit doubles the storage. This is a cost of full statevector simulation, not proof that any n-qubit hardware outperforms a classical computer.
Check yourself: why can a probability histogram miss a difference?
It records squared amplitude magnitudes and loses relative phase. For example, |+⟩ and |−⟩ have identical probabilities in the computational basis, yet H maps them to different basis states.
Gates change amplitudes.
X exchanges |0⟩ and |1⟩. Z keeps |0⟩ unchanged and multiplies the |1⟩ amplitude by −1. H maps |0⟩ to |+⟩ and |1⟩ to |−⟩. Applying H twice gives the initial state back. Gate order matters because these operations generally do not commute.
Starting from |0⟩: H → (|0⟩ + |1⟩)/√2 H → H → |0⟩ H → Z → (|0⟩ − |1⟩)/√2 H → Z → H → |1⟩
The second H makes the relative phase change visible in the final measurement. S and T apply phase factors i and exp(iπ/4), respectively, to the |1⟩ amplitude. They preserve probabilities at that instant but can affect later interference.
In the new workspace, bit order is q0 q1, with q0 on the left. For CNOT, choose distinct control and target qubits. The target flips only when the control is 1. In the existing lesson lab, follow its displayed target labels and probability order; do not transfer an SDK's bit-order convention without checking it.
Try it: compare H → Z with Z → H on |0⟩.
H → Z produces |−⟩; Z → H produces |+⟩. Their immediate measurement probabilities match. Apply H afterward to distinguish them.
Predict a distribution, then observe samples.
A computational-basis measurement returns a classical outcome and leaves the system in the corresponding measured basis state. Measuring |+⟩ returns 0 or 1 with equal probability. Applying H after that measurement does not recover the original coherent interference experiment.
A shot is one execution and measurement. In 100 shots of a 50/50 distribution, 50 of each outcome is possible but not required. Finite samples fluctuate. More shots improve statistical precision; they do not automatically remove device noise or correct a badly chosen experiment.
Write your expected probabilities before running. Afterward, compare the counts with those expectations and identify whether differences could come from sampling, noise, an incorrect circuit or a mismatch in qubit order.
Check yourself: do 54 zeroes and 46 ones disprove a 50/50 model?
No. Such a finite-sample result is compatible with a 50/50 distribution. Statistical evidence requires a defined test and sufficient data.
Prepare and inspect a Bell state.
Start at |00⟩. Apply H to q0, then CNOT with q0 controlling q1. The result is (|00⟩ + |11⟩)/√2. Both qubits are now entangled: the joint pure state cannot be written as a product of two individual pure states. Computational-basis measurements produce 00 and 11 with equal probability.
Apply Z to q0 to change the relative sign. Apply X to q1 to exchange the supported outcomes with 01 and 10. Combining these operations produces the four Bell states. Their phase matters even when measurement probabilities match.
H(q0) → CNOT(q0 → q1) → (|00⟩ + |11⟩)/√2 H(q0) → CNOT(q0 → q1) → Z(q0) → (|00⟩ − |11⟩)/√2 H(q0) → CNOT(q0 → q1) → X(q1) → (|01⟩ + |10⟩)/√2 H(q0) → CNOT(q0 → q1) → X(q1) → Z(q0) → (|01⟩ − |10⟩)/√2
Correlations alone do not let a party control a distant measurement outcome or send a faster-than-light message. A 00/11 histogram alone also cannot distinguish an entangled coherent state from a classical mixture. Phase-sensitive measurements are needed for that distinction.
Project practice: prepare a Bell state with a minus sign.
Use the circuit above, predict all four probabilities, and explain why an assessment should check the statevector as well as the probabilities. The independent assessment assigns one of the four target states and checks equality up to global phase.
Know what your experiment establishes.
An ideal simulation calculates a mathematical model. A noisy simulation adds chosen error assumptions. Real quantum hardware produces measurements from a physical device. Keep those labels with every exported result.
A real hardware record needs its provider task ID, device, circuit, shots, timestamps and measured output. The workbench retains these when the provider returns a completed task. Submission or a queued task is not a completed experiment. Physical noise can affect output even when a simulator predicts perfect behavior.
For a work problem, first define the objective, constraints, dataset and success metric. Choose a credible classical baseline. Compare solution quality, runtime and total cost, including preprocessing, state preparation, queueing and measurement. A small Bell experiment teaches a mechanism; it does not demonstrate an industrial speedup.
When reading a paper, separate its claim, experimental evidence, assumptions and your interpretation. Check quoted passages and equations against the source. A source-matched quote can still be misinterpreted by an AI model. A suggested teaching experiment is not a reproduced research result.
Investigate a paper →A clear standard for achievement.
The QuantumGridOS-issued Quantum Developer Foundations certificate demonstrates assessed foundational concepts and a small simulated circuit project. It is not an accredited degree, professional licence, hardware qualification or employment guarantee.
- Study this guide and practice with the linked labs. Reading progress alone never issues a credential.
- Complete 12 questions within 60 minutes. Score at least 80%.
- All critical concepts must be correct: amplitude versus probability, reversible interference, measurement, evidence of advantage, no-signalling and real-hardware provenance.
- Build the assigned two-qubit Bell state. The server checks the full state up to global phase, all four predicted probabilities, and your evidence and baseline interpretation.
- Work independently. Do not use the tutor, outside assistance or this guide during an assessed attempt. This is an unproctored assessment with an independence declaration.
- Up to three attempts per 24 hours. Unsubmitted attempts expire after 60 minutes. Review the concepts before retrying.
- Full access is required to take the assessment. Certificates are private by default, downloadable as SVG, and publicly verifiable only when you enable sharing.
- The credential records the assessment version and date. It does not expire; revocation is visible on a shared verification page. You can disable sharing or revoke your credential from the workspace.
Keep learning.
The next curriculum additions include advanced algorithms, noise and control, hybrid applications and specialist tracks. Your foundations credential describes the work assessed today.
Scientific references: IBM Quantum Learning and Amazon Braket documentation.