Personal
The rest of me.
The games I sank years into, the puzzles I couldn't leave alone, the people I learn from, and the stretches of my life that shaped how I work.
Most of how I approach engineering, I learned somewhere that wasn't a classroom. Usually it was a game, at 2am, with a friend on a call saying "wait, try that again."
The times of my life
- Growing upCall of Duty ZombiesWorld at War, Black Ops, Black Ops 2. Running Easter eggs with friends, notebooks full of symbols, and the first puzzles I ever obsessed over.
- Middle schoolMinecraft, and CosmicWriting small mods, then living on Cosmic's faction and prison servers, sometimes eight hours a day.
- High schoolWorld of Warcraft and CS:GO skinsPushing Mythic+ keys in Battle for Azeroth until I was top 30 Demon Hunter in NA for dungeons, and discovering that a video game's item market had arbitrage in it.
- 2022Computer science and finance at TMUThe puzzles got bigger and started paying rent.
- 2023 to 2025Nokia, then IBMLearning how real systems ship, and how to work inside teams much bigger than me.
- 2025 to nowSolana, Print.World, and my own bookNights and weekends on transaction landing and real-time models, a crypto trading terminal by day, and a systematic book I research and run myself.
Call of Duty Zombies, and the Easter eggs
Before I wrote any code, I was solving Treyarch's puzzles.
Zombies is a co-op mode in Call of Duty where you and up to three friends hold out against endless waves. Underneath that, Treyarch hid Easter eggs: long, multi-step quests with no tutorial, no quest log, and usually no hint that they existed at all. Finishing one unlocked things the normal game never gave you, like new weapons, upgrades, and mechanics, and moved the story forward. And the story was barely told on screen. It was scattered across radio messages, symbols on walls, and lines of dialogue you'd only hear if you did something strange.
Nobody solved these alone. When a new map came out, the whole community split the work: people dug through the game files for unused sounds and text, transcribed every radio message, collected theories in long forum threads, and tested steps live on stream until one of them stuck. Whole steps were found because one person noticed a sound cue that everyone else had walked past. Some of the hidden ciphers took the community years, and a few were never solved.
Knowing the steps was only half of it. Most of them needed all four players doing different things at the same time, in the right order, often on a timer, while still surviving the waves. One mistake usually meant starting the whole run over.
What it gave me: a love of systems that reward you for understanding them completely, and the habit of coordinating a group around a sequence where one person's mistake resets everyone. That's not far from a production deploy.
Minecraft, and eight-hour days on Cosmic
In middle school, Minecraft was everything. I started writing small mods, and a few of them ended up in modpacks other people played, which was my first taste of strangers running my code.
But most of my time went into Cosmic, one of the big competitive server networks, and specifically its factions and prison servers.
- Prisons are an economy game. You mine to earn in-game money, rank up to better mines, and trade with other players. The best players weren't the ones who mined the most; they were the ones who understood the economy: what was scarce, what people overpaid for, and when.
- Factions is team warfare. You build a base, defend it, and raid others. At Cosmic's level, that meant TNT cannons built with redstone engineering precise enough to break through layered walls, and bases designed around how those cannons worked.
I'd grind on those servers for eight hours a day. Looking back, it was my first exposure to two things I still care about: markets that reward understanding over effort, and engineering where the details are the whole game.
World of Warcraft: Mythic+ in Battle for Azeroth
In high school I got locked in on World of Warcraft during the Battle for Azeroth expansion, and specifically on Mythic+: dungeons with a timer, where every level up adds more health and damage to the enemies and new modifiers that change how you have to play.
At its peak I was a top 30 Demon Hunter in North America for Mythic+.
The interesting part of high-level Mythic+ isn't reflexes, it's planning:
- Routing. Every dungeon has more enemies than you need to kill. Choosing which packs to pull, in what order, and which to skip is an optimization problem with a timer on it.
- Cooldown planning. Big abilities have long cooldowns, so you plan which pull each one goes into before you enter the dungeon.
- Modifiers. The weekly modifiers change what's dangerous, so the plan changes every week.
- Small margins. At the top, a run is won or lost on seconds, and a single mistake in a pull can cost the key.
It's the closest thing I've done to latency work before latency work: you have a budget, you know where it goes, and you shave the slow parts one pull at a time.
CS:GO skins, and the first market I traded
The skin market deserves its own post, so it has one: trading the gap in CS:GO skins. Short version: it's where I first noticed that the price people see and the price they can get are different numbers, and where I first wrote code to trade the difference.
People I learn from
A few people whose thinking shows up in how I work:
- Corey Hoffstein (Newfound Research, Flirting with Models). His idea of rebalance timing luck changed how I look at every backtest: how much of the result is just a choice I made without a reason?
- Giuseppe Paleologo ("Gappy"). The way I think about risk: decide the drawdown you can live with first, scale exposure rather than switching it on and off, and stay in the game long enough for a modest edge to matter.
- Robert Carver. The most practical writing I know on sizing and forecasts in systematic trading, and on not fooling yourself.
- Marcos López de Prado. The deflated Sharpe ratio and backtest overfitting, which is why every result I trust has been corrected for how many things I tried.
What I've been reading
- Systematic tradingSystematic Trading, Robert CarverSizing, forecasts, and a framework for not overfitting.
- Trend followingWay of the Turtle, Curtis Faith, and The Trend Following Bible, Michael CovelWhy simple rules, held with discipline, get paid.
- ResearchAdvances in Financial Machine Learning, Marcos López de PradoBacktest overfitting, the deflated Sharpe ratio, and labeling done honestly.
- Portfolio managementAdvanced Portfolio Management and The Elements of Quantitative Investing, Giuseppe PaleologoRisk first, and portfolio construction as an engineering problem.
- Market structureTrading and Exchanges, Larry HarrisHow markets actually work underneath the price.
On the shelf right now
Most of my reading happens as audiobooks, on walks and commutes. It's mostly the history of markets and the people who built them: central bankers, hedge funds, Buffett, Simons, Thorp, and the spectacular failures in between, with a few books on how people actually think and decide.
What I watch and listen to
Most of what I learn outside of work comes in through my headphones.
- Atrioc on YouTube. Macroeconomics, markets, and general news, explained in a way that's genuinely fun to watch. It's my default for understanding why something in the news actually matters.
- Flirting with Models, Corey Hoffstein's podcast with systematic and quantitative investors. Long, technical conversations about how real practitioners build and size strategies, from the same person whose idea of rebalance timing luck shapes how I backtest.
- Validated, hosted by Austin Federa. Deep dives into how Solana actually works under the hood, with the people building it.
- Lightspeed from Blockworks. Another Solana show that goes into the intricate parts of the distributed system: how blocks get produced, how transactions land, and the trade-offs that let the network run as fast as it does.
The Solana shows in particular are where a lot of the ideas behind my transaction-landing work first came from. It's one thing to read about stake-weighted QoS or block engines, and another to hear the people who designed them argue about the trade-offs.
Small things
- I still like puzzles more than almost anything. If a system has a hidden rule, I want to find it.
- Most of my best ideas come from reading something unrelated and noticing it rhymes with a problem I have.
- I'd rather understand one thing completely than ten things roughly.
The original 1993 Doom, shareware episode and all, running right here in your browser.