About
What this blog is, and who writes it.
This blog
The blog is called Inference. It is where I write about building software in the age of AI, and it circles one question: what makes an intelligent system worth trusting. Everything lands in one feed, Inference, numbered in order of publication. There are four kinds of pieces:
Essays: an argument, made carefully.
Notes: how something works, in practice.
Fragments: one thought, said once.
Definitions: a term, pinned down.
What I am doing right now is on Now.
Who writes it
I am Edmo Lima. I build Dataqore, a context layer: it connects the systems a company already runs and turns them into answers that people, and now AI agents, can trust. I designed the whole thing before writing a line of code, because the mistakes that cost you are the ones you make at the start.
Before that, years building high-availability systems in fintech and adtech, at Nubank, Creditas, and OneFootball, among others. The kind that has to stay up and stay correct under millions of requests, where scale and resilience are the actual problem, not an afterthought. It taught me the thing that is easy to forget: getting software to work once is not the hard part. Keeping it correct when everything around it fails is.
On the side, nanocompile and a book on building software you can trust. The full story (bio, work, how to reach me) is on my main site, edmolima.com.
Elsewhere
The main site is edmolima.com. I am on X, and the code is on GitHub. Corrections and arguments are welcome.