Home Activity Blog Project Euler

About Me

I'm a Machine Learning engineer working for Klarna in Stockholm, Sweden. I work in the AI & Automation group, where we are tackling the problem of web automation at scale.

Prior to this, I completed my PhD in Astrophysics in 2018. I worked on a wide range of topics in astronomy, machine learning and statistics.


Contact

Feel free to email me
jimbarrett27@gmail.com

Or find me via my various online presences below

It's Jim's life, but not as we know it!

07/06/2016 - Jim Barrett

This morning I participated in the Astronomy and Astrophysics group Science Jamboree 2016, where I managed to win a prize (chocolates!) for the best talk. I'm not quite sure how I managed it, since in the midst of a series of informative and engaging talks, I performed a silly little poem;

For my science Jamboree
I'll tell you a bit about me
My journey through physics
in a few clumsy limericks
Right through to my PhD

I was born in July '91
In the blistering east Norfolk sun
And... oh wait a minute,
there's a strict time limit
So I had best be getting along

We'll skip ahead 18 odd years
Whilst I, still wet 'round the ears
kicked off my degree,
eager to see,
How I felt after 2 dozen beers

But I did learn some physics as well
Einsten, Newton, Maxwell
But as some of you know,
not a lot of astro
Although I think you can probably all tell

To be honest I found space a bit dreary
I preferred condensed matter theory
my vague proclivity
for superconductivity
"I should do a PhD in this clearly!"

Alas, I left those low temperatures
And became a software developer
but the real world is crap
I resolved to come back
And I somehow became an Astronomer....

My first project involved an obsession
with continuous autoregression
take an impulse response
to stochastic nonsense
et voila! banded matrix compression!

And with that O(n) inference
Time to write a paper of significance!
Except instabilities
destroy all utility
Try selling that at a conference!

So my PhD project, take 2!
But what was I going to do?
I felt it was prudent
to become Ilya's student
since bless him, he's only got a few

Binary population synthesis;
i.e how did it end up like this?
which hyperparameters
make a chirp that glamorous?
The combinations seem virtually limitless

So here's where I come in
When parameter space coverage is thin
we can interpolate
perhaps extrapolate
what happens when things spiral in

Now this work isn't finished yet
Ive tried random forests and neural nets
but despite all my labour
it seems nearest neighbours
may well be our safest bet

To conclude my science Jamboree
I'll make a short summary
my project got switched
because CARMA's a bitch
and I write really shit poetry!