Saksham Gupta
I like to understand things from the layer underneath.
When something doesn't make sense, I keep going down until it does: from the code to the runtime, and from the runtime to the machine.
Usually that means rebuilding it. I've written a 32-bit UNIX-like kernel, an out-of-order CPU simulator, and a few parsers, mostly to find out how the real ones work.
Right now I'm doing the same with distributed systems: reading the classic papers, and the ones behind today's large-scale storage and stream-processing systems, and writing up what I find here.