Python makes it easy to build APIs quickly, but many APIs that start clean become difficult to change as systems grow. New features, additional clients, and evolving requirements expose early design decisions that were never meant to last. This talk focuses on how to design Python APIs that remain understandable, adaptable, and safe to modify over time.
The session draws from real backend systems built and maintained in production. Rather than presenting framework tutorials, it focuses on design choices and trade-offs that repeatedly affect long-term maintainability.
The talk begins by clarifying what maintainability means for APIs in practice. Beyond readable code, maintainable APIs allow change without fear, make behavior easy to reason about, and reduce the risk of unintended breakage. Common warning signs of declining maintainability will be discussed to help teams recognize problems early.
Next, the talk examines API boundaries and responsibility separation. It covers how unclear boundaries between routing, business logic, and data access lead to fragile systems, and how clearer structure supports growth without unnecessary complexity.
Validation and contracts are then addressed as core design concerns. Treating requests and responses as explicit contracts helps prevent bugs and misunderstandings between systems. Practical validation approaches that worked in production will be shared, alongside examples of shortcuts that caused long-term issues.
The session also covers authentication and authorization design, with attention to where security logic belongs and how poor placement can limit future change. Patterns that allow permission rules to evolve safely will be discussed.
Finally, the talk looks at API evolution and change management. It explores when versioning is necessary, when it signals deeper design problems, and how APIs can evolve without breaking clients. The session concludes with lessons from real failures and what could have been done differently.
This talk is intended for developers with experience building Python APIs who want to move beyond basic functionality and design systems that remain maintainable as they scale.
