Selected work

Systems shipped under real constraints.

Public launches and owned products built to survive real users, volatile markets, and difficult infrastructure.

02 Ethereum · Smart contracts · 2022
Yakuza Inc. number 1 red pixel-art character from the Ethereum collection

Lead smart contract and launch engineer

Yakuza Inc.

Lead engineer for a seven-figure Ethereum mint: contract, allowlist, sale controls, metadata, and payout path, carried from deployment through a 3,223-token sell-out.

A seven-figure, 3,223-token mint that sold out in under two hours; the final 596 public mints landed in 56 seconds.
03 On-chain game · Bitcoin · 2024–25
Pizza Pets pixel-art wordmark for the Bitcoin on-chain game

Product engineering

Pizza Pets

Product engineering for a fully on-chain Bitcoin game where more than 58,000 pets can be fed, evolve through six stages, or die permanently—without an off-chain game server.

More than 58,000 pets and one million on-chain interactions turned recursive inscriptions into a living Bitcoin product.
04 Permanent archive · Bitcoin · 2024
Project Spartacus I Am Assange archive wordmark

Inscription pipeline engineering

Project Spartacus

Inscription-pipeline engineering for one of the largest public-interest Ordinals runs: mapping 76,911 Afghan War Diary records into a resumable, verifiable multi-block archive on Bitcoin.

A pipeline built for 76,911 public records—one of the most operationally demanding Ordinals archives attempted.
05 AI product · Publishing · 2026
Agentic Brief mobile interface showing a personalized technology news signup

Publisher and builder

Agentic Brief

An owned AI publishing system that turns a reader’s stated interests into a confirmed, structured preference profile before any edition is sent.

A production double-opt-in and preference pipeline—not a form that drops unverified addresses into a list.

The standard behind the work

Different systems. The same engineering discipline.

Across every project, I stay close to the critical path, plan for failure, and leave clear evidence behind.

01

Close to the critical path

I stay with the part that can sink the product, not only the part that is easy to present.

02

Ready for failure

Retries, limits, review states, and operator visibility are part of the system from the start.

03

Evidence left behind

Public sources, readable architecture, instrumentation, and a clean handoff make the work easier to trust.

A different system. The same standard.

Let’s talk about what has to hold up.

Start a conversation