It's easy to design software on fast internet.
It's much harder to design software for reality.
While working on a learning platform serving students across multiple African countries, bandwidth wasn't an edge case. It was the environment.
Students accessed the platform using mobile data. Connections were inconsistent. Network costs mattered. Every unnecessary request had a real financial cost attached to it.
That changes how you think. Suddenly performance isn't a nice-to-have. It's part of the product.
Database design becomes a product decision. Caching becomes a product decision. Payload size becomes a product decision.
The most elegant architecture in the world means very little if users can't comfortably access the product.
Working under these constraints taught me something valuable: Good engineering isn't building for ideal conditions. It's building for the conditions users actually experience.
The closer engineering gets to reality, the more valuable it becomes.
Constraints are not obstacles. They're design inputs. The best systems acknowledge them instead of pretending they don't exist.