Tasks
Task #1
Explain on a high level how do you envision the architecture of a Kotlin compiler fuzzer?
As you will be using the result of this task in other problems, feel free to return and refine and reiterate the design. If possible, track your increments and also present them in your answers.
Important: it is OK to use AI during your work on these tasks, but it is important for you to also record and explain in your answers what you did to ensure the process of working with AI gets you a good solution.
Task #2
Given the architecture from task 1, which of its components are the most important to support pluggability? Why?
Task #3
Given the architecture from task 1, do you see the potential of using LLMs / AI for some of its components? How?
Task #4
How would you test the Kotlin compiler fuzzer itself? Is it possible to back-test the fuzzer? How?
Task #5
(Bonus) Given the architecture from task 1, select one of its components and explain how it could be implemented in more detail. For example, what existing tools or libraries could be used there? Or which algorithms could be used to implement that component?