research fellow, iit kanpuri work with prof. adithya vadapalli in cse at iit kanpur as a research fellow. the problem is secure computation: several parties should be able to evaluate a function over private inputs and learn only what the protocol allows.
that puts me around secret sharing, garbled circuits, oblivious transfer, adversary models, and the communication costs that decide whether a construction can leave the paper. round count matters. so does the exact point where a party can abort or infer something from timing.
i like the part where a clean proof meets a real network. the protocol still has messages, delays, malformed inputs, and machines that disappear at the worst possible moment. i am 16, still in high school, and somehow this is where most of my reading goes.