M L L . Computer

I joined Zora in 2021 and worked across its evolution from cryptomedia marketplace, to onchain social network, to attention market.

The core bet stayed consistent: public access and creator ownership do not need to trade off. Media can be free to see, circulate, and participate in, while still having value, provenance, and ownership.

My work focused on the creator surfaces that carried through those shifts.

Creator tools

I led engineering on Zora’s early creator tools and frontend media surfaces, including the create flow, multimedia rendering, text posts, and onchain custom themes.

I co-led the team behind canvas posts: user-authored spatial compositions made from images, text, and shapes.

These tools were small on the surface, but central to the product: they shaped how creators understood what they were making, how it would appear, and what it meant to publish something onchain.

Protocol and media systems

I built JsonExtensionRegistry, a small onchain extension contract for attaching arbitrary JSON metadata to other contracts. It was roughly forty lines, and still does what it needs to do.

I also designed parts of Zora’s decentralised media stack: upload, processing, CDN pipelines, IPFS permanence, and onchain themes. The goal was simple: what creators publish should not live or die with the platform.

New primitives and experiments

I architected Zora live streaming: Cloudflare ingestion, Livepeer archival, IPFS permanence, GraphQL realtime, and tokenised stream coining. The experiment asked what happens when attention to a live stream accrues to the streamer rather than the platform.

I shipped AI canvas, powered by fal.ai, with revenue routed through Splits to creators and model providers. It put AI directly inside a creator tool, while paying for inference as part of the creative and economic system rather than hiding it as a platform cost.

Engineering leadership

Alongside product and protocol work, I managed engineers, hired and sourced candidates, supported career development, handled performance processes, and led cross-functional delivery.

I also led internal AI enablement: contribution policy, AI code review process, design and marketing workshops, containerised frontend agent tooling, and spec-driven workflow experiments.

Before Zora, I spent years working on image collecting, spatial canvases, collaborative publishing, and tools for shared thinking.

Bench.li (2013–2017)

Bench.li was a design blog built around reference image collecting.

A custom Chrome extension turned right-click into post: an image could be captured with page context, uploaded, and published immediately.

It had a small, steady readership in design Twitter, and was later co-edited with Luke Robertson.

muse.li (2014–2017)

muse.li was a long-running attempt to understand spatial software as a thinking tool.

I rebuilt it many times: solo canvases, real-time collaboration, an Electron desktop client, drag-and-drop hypermedia, and early experiments in shared visual workspaces.

The recurring lesson was that spatial tools are easy to prototype and very hard to make habitable. Each version was a guess at what a shared thinking surface should feel like.

KNOW (2018–2019), ETH Zürich

KNOW began as a publishing tool for an ETH seminar and became a real-time collaborative canvas for forty students working across philosophy and architecture.

I built it with Norman Sieroka and Hannes Mayer for Philosophical Reflections on Digital Methods in Architecture: two disciplines thinking together without a shared knowledge format.

The project was on track to spin out through ETH, but its complexity eventually outgrew us. It was later documented in the ETH Learning and Teaching Journal by Sieroka and Mayer.

means.at (2020)

means.at was a peer-to-peer publishing project for six artists and two programmers, built on Beaker.

There was no central host. Sites could be forked like sentences in a conversation.

Where earlier projects assumed a place where the work lived, Means asked what happens when there isn’t one.

I co-directed Public Office, a Melbourne studio working between print and the web.

I designed and built bespoke publishing systems, CMSes, archives, and experimental web tools, usually as technical lead and sometimes as designer.

We worked with artists, galleries, journals, architects, publishers, festivals, and people whose work did not fit neatly into any of those categories.

Most projects began with a publishing problem that was not quite served by an off-the-shelf CMS. The interesting decisions were rarely just technical. They were about structure: what the system should name, what it should preserve, and what it should refuse to formalise.

Alongside the studio, I taught and ran workshops at RMIT, and continued a long collaboration with Paul.

Selected projects follow;