Garden
Building large-scale, reliable systems is a multi-disciplinary endeavor. It is better to generally know the architecture landscape and different options available at each layer and dig into a specific topic when the need arises. You have to carefully balance practicality and efficiency after knowing alternatives.
Most ambitious systems tend to be data-intensive, and handling I/O is a crucial aspect of applications. Data is where you should begin your Ambitious journey. Start with understanding what Data Model options are available and which ones fit your use case.
Large systems tend to get associated with NoSQL, schemaless databases by default, but that is not generally the right choice. You should know the advantages of Transactions and architect your application to protect data sanctity, irrespective of the database you use.
Applications also need to deal with Domain Complexity to keep up with the business. Domain-driven Design has proven to be an effective strategy in keeping complexity low as systems evolve.