The best software engineering practice for a solo developer is the one that automates the roles you cannot hire for. You are the PM, the QA, the DevOps, and the support team. The methodology has to absorb that.
A solo developer is not a developer with a smaller team. A solo developer is the entire engineering organization compressed into one person: product manager, quality assurance, DevOps engineer, release manager, incident responder, and developer. Five roles. One brain. One day.
The way solo devs fail is not by writing bad code. They fail by drowning in the roles they didn't hire out — silently doing PM work in their head, deferring QA until the customer finds the bug, treating "deployment" as something they'll set up "soon." The compound interest on that deferred work is what kills the project at the 90% mark.
The discipline that survives this is one that automates each role you cannot hire for. CI handles QA. Feature flags handle release management. Observability handles incident response. The daily ship handles project management. SSD is the name for the integrated form of that discipline.
Before the first business logic exists, "Hello World" is in production. The deploy pipeline is solved before anything else. If I cannot deploy, I do not have a product — I have a research project.
Every day I work on this project, something ships. A bug fix, a feature behind a flag, a performance improvement, a documentation update — but something ships. A day without a ship is a process failure.
I cannot QA my own code in my own head. CI runs the tests, the linter, the type checker, and any production smoke checks. If CI is green and the feature flag is off, I sleep at night. If CI is red, the system is broken.
No long-lived branches. Everything lands on main, behind a feature flag, off by default. The code lives in production from the moment it's written. The user sees it when I flip the switch — not before.
The Nightly Ritual: tests pass, code committed and pushed, CI green, flags set, tomorrow's first task identified. Future me should be able to pick up cold and produce value in fifteen minutes.
A concrete starting point. Not the only stack — just one that works on day one for a single person:
Read the full SSD methodology — designed for the team of one.
Read SSD