What core principles guide a new AWS cloud for scalable SaaS?

[OliviaM]

As a lead developer at a startup, I'm tasked with designing our first serious cloud computing architecture on AWS to support a new, scalable SaaS product, moving us beyond a single monolithic server. I'm comfortable with individual services but feel overwhelmed by the sheer number of interconnected decisions around networking, security groups, database choices, container orchestration, and cost optimization that will have long-term consequences. For architects who have built greenfield cloud platforms, what foundational principles or design patterns did you establish first to ensure scalability and maintainability? How did you approach the trade-offs between managed services versus self-managed infrastructure for control versus operational overhead, and what tools or frameworks did you use for infrastructure as code and monitoring from day one to avoid technical debt?