Learn quantum computing
by collapsing it yourself.
A self-paced path from "what is a qubit" to running real circuits on AWS Braket — built for engineers, researchers, students, and leaders who don't yet speak the math.
What brings you here?
Pick the description closest to you. It won't lock you out of anything below — it just orders the same seven modules to match your starting point.
The four ideas everything else is built on
Skip the physics department. These four concepts are the entire foundation — everything from Shor's algorithm to a Braket pricing page traces back to them.
The unit of quantum information
A classical bit is 0 or 1. A qubit is described by two numbers (amplitudes) that determine the probability of reading 0 or 1 when measured — and, crucially, it can hold both possibilities at once before that measurement happens.
Both states, weighted
A qubit in superposition isn't "secretly" 0 or 1 waiting to be revealed — it genuinely occupies a combination of both. The hero coin above is a superposition until you click it; clicking forces a measurement and the state collapses to one outcome.
Correlated, not connected by a wire
Two entangled qubits share a single description. Measuring one instantly determines what the other will read, no matter how far apart they are — a correlation with no classical equivalent. Try it in the circuit lab below.
The one-way door
Reading a qubit's value ends its superposition — you get one classical bit out, and the rest of the quantum information is gone. Quantum algorithms are engineered so that, by the time you measure, the *wrong* answers have cancelled out.
Build a one-qubit circuit
Apply real quantum gates to a single qubit and watch its amplitudes change. This is a genuine (if simplified) state-vector simulation, the same idea Braket's SV1 simulator runs at much larger scale.
Now entangle two qubits
This applies a fixed circuit — Hadamard on qubit A, then CNOT from A to B — producing the Bell state (|00⟩ + |11⟩)/√2. Measure it a few times: the two results always match, even though each one is individually random.
What quantum computers are actually good for
Not "faster at everything" — faster at a specific, narrow set of problems where quantum interference can cancel out wrong answers. These five come up constantly, including on Braket.
Grover's search quadratic
Searches an unsorted list of N items in roughly √N steps instead of N. Useful wherever you're checking candidates against a condition — cryptanalysis, database search, optimization subroutines.
Shor's factoring exponential
Factors large numbers exponentially faster than the best known classical methods. The reason RSA encryption is considered vulnerable to a large enough fault-tolerant quantum computer, and why post-quantum cryptography is now an active field.
VQE hybrid
Variational Quantum Eigensolver — estimates the lowest energy state of a molecule, alternating between a quantum circuit and a classical optimizer. The leading near-term approach for quantum chemistry and materials design.
QAOA hybrid
Quantum Approximate Optimization Algorithm — tackles combinatorial optimization (routing, scheduling, portfolio selection) by encoding the problem as an energy landscape and searching for low-energy configurations.
Quantum machine learning exploratory
Uses parameterized quantum circuits as trainable models, or quantum kernels inside classical ML pipelines. Still early and unproven at scale, but one of the most active research areas today.
Quantum simulation native
Using a quantum system to simulate another quantum system — chemistry, materials, condensed matter — was the original motivation for quantum computing, and remains its clearest long-term use case.
From simulator to real quantum hardware, one SDK
Amazon Braket is a fully managed service for building, testing, and running quantum circuits — on classical simulators for development, and on real quantum processors from independent hardware providers, all through one Python SDK and one console.
Devices available on Braket
| Device | Type | Technology | Good for |
|---|---|---|---|
| SIMSV1 | State-vector simulator | Exact simulation, up to ~34 qubits | Developing and debugging circuits before spending on hardware |
| SIMDM1 | Density-matrix simulator | Models noise and decoherence | Predicting how a circuit behaves on imperfect real hardware |
| SIMTN1 | Tensor-network simulator | Approximate, structure-exploiting | Larger, sparser circuits that exceed SV1's qubit limit |
| QPUIonQ | Trapped-ion processor | Ytterbium ions, all-to-all connectivity | High gate fidelity, algorithms needing flexible qubit connections |
| QPURigetti Ankaa-3 | Superconducting processor | Tunable superconducting qubits | Fast gate times, larger qubit counts |
| QPUIQM Garnet | Superconducting processor | Square-lattice, tunable couplers | EU-hosted workloads with data residency needs |
| QPUQuEra Aquila | Neutral-atom processor | Rubidium atoms, analog + digital | Direct Hamiltonian simulation of physical systems |
Hardware providers on Braket change over time — always confirm the current list and Region availability in the AWS Braket console before building on a specific device.
Your first circuit: a Bell state on Braket
from braket.circuits import Circuit from braket.devices import LocalSimulator # Build: Hadamard on qubit 0, then CNOT(0 -> 1) circuit = Circuit().h(0).cnot(0, 1) # Run locally first — free, instant, no hardware needed device = LocalSimulator() result = device.run(circuit, shots=1000).result() print(result.measurement_counts) # {'00': ~500, '11': ~500} — the Bell state you built in the lab above
from braket.aws import AwsDevice device = AwsDevice("arn:aws:braket:us-east-1::device/qpu/ionq/Aria-1") task = device.run(circuit, shots=1000) result = task.result()
How a Braket workflow usually goes
Pricing is task- and shot-based and varies by provider and Region — check the current Amazon Braket pricing page before running anything at scale, since simulator and QPU rates are updated by AWS independently of this guide.
Where people actually land
Quantum computing isn't one job — it spans hardware physics, algorithm research, software engineering, and go-to-market. Here's where technical and non-technical backgrounds tend to fit.
Quantum software engineer
- Python, linear algebra basics
- Circuit-level SDKs (Braket, Qiskit, Cirq)
- Classical software engineering discipline
- Braket SDK and Hybrid Jobs
- Building reusable circuit libraries
Quantum algorithms researcher
- Quantum mechanics, complexity theory
- Designing and proving new algorithms
- Reading and publishing papers
- AWS Center for Quantum Computing collaborations
- Braket for empirical validation of new methods
Applications / domain scientist
- Deep domain expertise (chemistry, logistics, risk)
- Translating problems into VQE/QAOA formulations
- Comfortable being a "translator" between fields
- Hybrid Jobs for domain-specific pipelines
- Working with PennyLane / Ocean plugins on Braket
Quantum solutions architect
- Strong AWS fundamentals (IAM, S3, EC2)
- Enough quantum literacy to scope real pilots
- Customer-facing communication
- Braket architecture and cost design for customers
- Bridging classical HPC with quantum workloads
Product / program / technical strategy
- Working literacy in the concepts on this page
- Reading a landscape of hardware providers and timelines
- Translating technical risk into roadmaps and budget
- Scoping when Braket pilots make sense vs. waiting
- Vendor and provider evaluation
Glossary
The terms you'll hit repeatedly, in plain language.
Check yourself
Five questions covering everything above. No pressure — wrong answers just show the explanation.