About Unique Divine

I’m a programmer that makes music and writes.

I’m obsessed with learning, solving problems, and creating, whether this involves artwork, music, coding, anything. To learn and to create is my ikigai (生き甲斐).

What I do for work

I’m a Co-Founder and one of the lead software engineers for a Web3 project called Nibiru. Formally, I’m the CEO of tech company called Nibi Inc.. I’m also an open-source contributor. I build mostly in the artificial intelligence and Web3 spaces.

What I’m doing right now

In learning, there are times for exploration and times for exploitation. That is, there are times to expand your knowledge base to get more source material and times when you should make use of what knowledge is already there, leverage the sharpened saw, and finish what is started.

I am in this second mode. Lots of programming. Lots of strategizing. No new books or entertainment. Just growth and execution.

Contact

The fastest way to get in contact with me is probably a direct message on X/Twitter, but e-mail works too.

Passions and interests

• 🧠 Anki: I’m a huge nerd about spaced repetition software and other strategies for accelerated learning. I’ve been using Anki to learn faster and remember more of what I learn for probably 5+ years at this point. Although I have over 100,000 mature cards, the reviews take up less than 10 minutes a day when I stop adding new material. Using Anki effectively was a crucial element in my journey to learn Japanese to fluency.

• 🎼 Music: I try not to go a day without practicing 🎷 or 🎹. Though this isn’t always possible with things like travel, music forms an essential part of my life.

• 🔥 Coding: My drug of choice.

• Physics/Sciences : Up until the age of 23, almost all of my time went into developing competencies in physics, applied mathematics, and machine learning.

  • I was almost certain I wanted to become a physicist as I went throught university but ended up more excited about other things. Although my professional trajectory has evolved a lot since then, I’m still fascinated by physics and mathematics, and I find myself going down rabbit holes and digging into new publications more than I’d like to admit.
  • Exploring what truths are fundamental to reality and building an intution about these principles pushes the right buttons for me.

• 📚 Reading: Some time in 2021, I decided to read and take notes on roughly two books a month. I mostly read biographies, self-help, and books on science and tech. Some of my recent favorites were:

  • About Face: The Odyssey of an American Warrior by Colonel David H. Hackworth
  • Something Deeply Hidden by Sean Carroll
  • Machines Who Think by Pamela McCorduck
  • Diaries of Adam & Eve by Mark Twain

My drug of choice (coding)

In the later part of college, I was working on research in gravitational lensing, this effect that’s like seeing the world through warped glass, except the glass is gravity itself.

Imagine that “empty” space isn’t empty. It’s more like soft, stretchy fabric. And if we put something massive on it like a star or a galaxy, the fabric bends. All things with mass are “massive”. It’s just that the bends are harder to notice when mass is small.

Objects have to follow the paths created by these bends. And light doesn’t get a pass. Even though it’s massless, it follows the curves of our gravitational fabric. So when light from a distant object, say, a galaxy far, far away, passes near something huge, its path bends.

We don’t see the bend. We just see the result. Things look smeared, magnified, even multiplied. A single galaxy might appear as arcs. A distant quasar might echo around a foreground mass like it’s glitching.

But there are no glitches in life. This lensing effect can be quantified, estimated, and more well understood.

It turns out, you can estimate the physical parameters from observed gravitational lensing with images of the distortions and deep neural networks. I chose to get better at coding in order to work on this research.

And by some turn of events, I realized the tools I was using to complete my research (AI predictive models) were more enthralling to me than the objectives to which I was applying them. That was where I shifted more from a physicist or mathematician to a technologist.

Coding aligns so closely with my natural inclinations. It has the perfect mixture of engineering, problems to solve as puzzles, and applied creativity. It’s like playing with legos as a kid again, except you can build anything.

I’ve written well over a million lines of code at this point and have no intention of stopping.

I’ve always believed that you have to go all in on something to find out if you’re really made for the craft or if you just like the sound of the end results.

Some open source code I’ve written:

Repository	Description
NibiruChain/nibiru	Nibiru source code. Nibiru is a breakthrough smart contract platform powering an ecosystem of decentralized applications.
NibiruChain/nibiru-wasm	Rust monorepo and smart contract prototyping hub for Nibiru. Includes dev tooling, production and example smart contracts, an end-to-end testing framework, and other useful libraries.
NibiruChain/ts-sdk	TypeScript-based client SDK for interacting with the Nibiru blockchain
💻 grid-world-plus	A research project in vision-based, deep reinforcement learning. The goal is to improve sample efficiency using transfer learning, unsupervised representation learning, and attention mechanisms to interpret memories.
🐍 python-crypto-api-wrappers	A Python package for wrapping common Cryptocurrency APIs such as Messari, Coin Market Cap, Etherscan, CoinGecko, etc.
📝 Commonplace book	Notes, tutorials, and reference materials for Python, Golang and the Cosmos SDK, Typescript, React, Artificial Intelligence, and effective learning. [book] [repo]
🧬 GANs-for-Genomics	Computational genomics research. I leverage deep learning to generate and learn from synthetic DNA samples.
🏫 Langevin-Dynamics-for-NN-Optimization	Bayesian Neural Network (BNN) implementations based on Langevin Dynamics and tested on real-world data
cellar-subgraphs	Subgraphs to process and store entities for blockchain events emitted from the Sommelier Cellar contracts. Built using The Graph’s AssemblyScript library, graph-ts.