Open Research Institute 🔗 ORCID: 0009-0001-0857-096X

I am a data engineer, computational neuroscientist and learning expert, I model how data gets shaped and stored as meaningful information. I've implemented a digital data structure based on how the brain samples and stores memory, and a full database and learning architecture from the ground up.

Currently working on building a digital mind, a language model with geometric embeddings instead of flat vectors approximations. Modern datacenters need around 10.000-20.000W of power to power current AI like Claude or GPT, while the human brain runs on 20W, so the aim is to close that gap.

This is the digital mind project, where my free time goes and research gets applied directly

What if you could access an expert at any field, instantly from your cellphone? What if you could find other researchers working on the same problems, and be found by others looking for your expertise?

To move science ahead, we put it all in one place, facilitate communication between researchers, and create a digital model of our world where agents can orient themselves without getting lost in hallucinations.

The backend underneath the Open Research Network: a geometric data structure modeled after how the brain samples and stores information, with a SQL-compatible database and cognitive architecture built on top — Turing complete and extremely efficient. Here are some of the benchmarks:

find the 1 corrupted record in a million      ~1.8× faster than a Merkle tree
find all 50 corrupted records in 100,000        50× faster than a Merkle tree
SQL queries, vs SQLite with indexes            7–10× faster
data-integrity check, vs SQLite            ~19,000× faster

Run them yourself and see the rest at →

My R&D repository holds the published papers and active work on a geometric description of thought and language. My home is in psychology — the first paper is essentially philosophy, from there it gets more mathematical.

Whitepapers →

Liminal Gravity → — my substack, where I write simply, with pictures and casual language, without worrying about academic pedantry.

All of this is independent work done in my own time and released for free. If you find it useful or want to see it continue, any support helps.

For collaboration or employment: walter.h057@gmail.com

• Former AI/crypto dev at Renlabs/Torus Network
• Former full stack dev at NETDigital, a small ISP company
• Published peer-reviewed neuroscience research
• Psychology student