What’s the best way to learn system architecture and design?

Think of architecture as a sequence of decisions, not as selecting a fashionable pattern. Write down the assumptions behind each decision and define how you will test them. Keep module-level tests for business behavior, contract or integration tests at important boundaries, and a small set of end-to-end tests for critical user journeys. When production or development reveals a problem, document which assumption was wrong and adjust the design. That feedback loop is more valuable than trying to create a perfect architecture upfront.

A useful study path is to combine broad principles with real system examples. Designing Data-Intensive Applications is excellent for storage, consistency, distributed systems, and trade-offs. Head First Design Patterns is a gentler introduction to reusable design ideas. The AWS Well-Architected Framework is also worth reading, even if you translate the product names into general concepts such as reliability, security, operations, performance, cost, and sustainability. Reading technical papers and architecture write-ups, then applying the ideas to your own project, will make the material stick.

Architecture is learned by repeatedly turning requirements into the simplest system that can satisfy them, then operating and improving what you built. Start with concrete constraints such as expected traffic, reliability, consistency, security, deployment frequency, budget, and who will maintain the system. Experience comes from building smaller pieces, seeing which assumptions fail, and applying those lessons to larger systems.