Daniel Whiteson

So we're interested in the rare stuff.

So we have a filter at the very, very early stage that decides keep it or kill it.

And that makes downstream analysis much more efficient because you don't have to search through all the boring stuff to find the interesting stuff.

But it means also we have to be smart about what we're keeping and what we're killing.

That's actually what my team works on.

And I found that super fun.

You have to make this super fast decision and you don't have a lot of time to do a lot of really fancy calculations.

Um, it's killer to keep it every 24 nanoseconds.

Right.

So high speed computing, I thought was a really fun challenge.

Um,

how do you know you're not throwing out the next Nobel Prize?

I mean, if we're talking, I'm assuming you're using machine learning or AI of some kind.

Well, the very first stage is very simple, and then it gets more complex, and we're definitely using machine learning and AI.

We don't know that we're not throwing away some treasure out with the garbage, but we do have some filters that just randomly select events.

Let's just keep one out of a thousand randomly, so that if there's something crazy that we didn't expect, we'll probably find it there.

But we just can't keep all of it.

It's just too much data.

We're talking about petabytes and petabytes every day.

It's insane how much data we produce.