Every blog starts with one of these, so here we are.
I'm Leon — founder & researcher at Empero, online as kodee. I spend most of my time running experiments on efficient language model architectures, and most of what we release ends up under the empero umbrella on github and huggingface.
This blog is where the write-ups live:
- experiment logs — things that worked, things that didn't, numbers included
- notes behind the papers (see the publications list on the home page)
- assorted opinions on training, serving, and the tools in between
if an experiment doesn't get written down, did it even happen
More soon. If you want to say hi, the guestbook is open.