practical system design w/ bassem dghaidi

Sourced from the interview between Patrick Akil and Bassem Dghaidi (GitHub).

Disclaimer: The following notes are an AI-generated summary based on the transcript of the video.

Vertical scaling is vastly underrated Before jumping to complex sharding or microservices, maximize vertical scaling. Modern nodes are powerful (hundreds of CPUs); GitHub handles millions of requests on surprisingly small, basic clusters.
Design only for the next order of magnitude Do not design for 10 years in the future. If you have 100 users, design for 1,000, not 100 million. Building for hypothetical “hyper-scale” creates unnecessary complexity and cost (e.g., the startup that almost failed by migrating to Kubernetes too early).
Software is evolved, not built Treat software like a living organism that fights entropy, not a static construction project. You must solve specific problems you have today rather than building generic frameworks for problems you might have tomorrow.
Simplicity is the ultimate sophistication at scale The easiest solution is complex enough. Avoid abstractions and edge cases. Write code in the “dumbest, most simple way possible.” Complex architectures look impressive on a resume but often result in a “fatter AWS bill” and slower feature shipments.

Professional engineering is not a hobby Drop the obsession with code aesthetics and “craftsmanship” when on the clock. Your job is to solve business problems. Sometimes “good enough for today” is the correct technical decision to give the business runway.
Translate technical debt into business risk Business stakeholders do not care about abstract refactoring. You must reframe technical requests in terms of their impact on revenue, customer experience, or operational risk to get buy-in.
Quantify your impact with numbers, not fluff To get promoted or recognized, you must prove your value. Don’t describe what you built; describe the measurable impact it had on the business (e.g., reduced latency by X%, increased user retention by Y%).

Adaptability is the new core competency With AI agents now capable of writing 90% of the code, the value of a human engineer shifts from syntax proficiency to operational oversight and the ability to learn new domains rapidly.
Shift focus from writing to operating As AI takes over the “grunt work” of coding, senior engineers must focus on reliability, security, and ensuring systems “never go down.” You become the architect and pilot, while AI is the mechanic.
Interview theory vs. production reality There is a massive gap between system design interviews (which are often “hacked” via memorized theory) and actual production engineering. Big Tech interviews often test for scale that most companies and engineers will never actually need.

nomomon