AI agents code explosion. Building and Testing massive codebases
With the rise of AI agents the size of our codebases is growing faster than ever before. The larger the codebase, the more time it takes to build and test which slows down our development velocity which the AI agents are meant to increase. Modern build systems such as Bazel are designed to allow remote execution which allows you to scale horizontally rather than vertically bringing your build speeds close to their theoretical minimum. This talk will discuss what remote build execution is, how it works and why it's the gold standard for building and testing.
Unnar Freyr Erlendsson
Unnar is an experienced software developer who has spent his career working in developer tooling. He started his career at Google Stadia working on a build and debugging plugin for game developers before moving to EngFlow where he is assisting some of the largest companies in the world build and test their codebases with remote execution.
He has also been instrumental in building up the competitive programming scene in Iceland and has been involved in writing problems and officiating in Forritunarkeppni Framhaldsskólanna since 2014. He is one of the founders of Keppnisforritunarfélag Íslands.
