00/Open engagements
Two seats for designers. One for whoever finishes our poker project.
We hire in whole problems rather than in tickets. Each of the three engagements below is an area someone owns outright, and each ends in a paid evaluation that we pay for whether or not we make an offer.
01/Engagement 01
Own our poker project, start to finish
We are looking for one exceptional engineer — or a small team — to take complete ownership of a specified, funded, unstaffed on-chain poker product and carry it to launch.
Our games squad is three engineers. They are also the people who understand our settlement contracts best, and they are committed through Q4 to migrating the distribution engine that sits underneath all forty-one properties and every edition we have issued. That migration cannot slip, and it cannot be handed to someone unfamiliar with the code.
Poker is genuinely important to us. It is not, however, the thing that breaks the company if it launches late. Given the choice between six months of divided attention and handing it to someone who can make it their entire focus, we would rather do the second.
This is not a rescue of a failing codebase. There is a forty-page specification, a reviewed architecture, and about three weeks of exploratory contract work. You would be starting from a well-lit position — which is exactly why we want to hand it over now rather than after a year of stop-start effort.
What you would build
Six components, roughly seven months
One very strong engineer can complete this in about seven months; a team of three in roughly half that. The first two components are where the genuine difficulty lives — the rest is careful, conventional engineering.
Mental poker protocol
Commutative-encryption shuffle with distributed key generation, so no single party — the house included — can see a card that has not been dealt to them.
Hand evaluation engine
Deterministic seven-card evaluator with exhaustive test vectors. Must produce identical rankings client-side and in the settlement contract.
Table server
Authoritative real-time server managing seats, blinds, action clocks, disconnect protection and side pots across multiple concurrent tables.
Settlement contracts
LIGO contracts for buy-in escrow, pot distribution, rake collection and a dispute path that can resolve a table if the server disappears mid-hand.
Client
A TypeScript web client with wallet connection, table view, action UI and full hand history — built against the visual system the Editions design team owns.
Anti-collusion & audit
Hand-history export, pattern detection hooks, and a published verification tool that lets any player re-derive the shuffle after a hand closes.
At a glance
- Engagement
- Full ownership — individual or team
- Shape
- Fixed-scope contract, or full-time
- Duration
- Approx. 7 months to launch
- Location
- Remote — US or ±5h of US Central
- Compensation
- $190k–$260k, or fixed bid
- Reports to
- CTO
What we genuinely need
- Production TypeScript at a level where you are comfortable owning a real-time server and its client alone.
- Demonstrable smart-contract work — LIGO or SmartPy on Tezos ideally, but strong Solidity or Rust experience transfers and we will fund the ramp-up.
- Applied cryptography literacy: you should be able to read a mental-poker paper and tell us where the protocol would break under a hostile player.
- Experience with authoritative real-time servers — the failure modes of reconnection, clock drift and out-of-order messages should be familiar ground.
- A track record of finishing. This project's defining risk is abandonment, not difficulty.
What would help, but is not a filter
- Prior card-room, casino or betting-exchange engineering experience.
- Formal verification exposure, or any background in proving properties of code.
- Familiarity with Taquito, Beacon or the wider Tezos tooling ecosystem.
- Regulated-gaming compliance experience in any jurisdiction.
Apply
No cover letter. Send the work — a repository, a paper, a portfolio, a plot, a contract you are proud of. Anything that shows how you think.
Process
Five steps, about three weeks
Tomás Guerrero runs all of it. You will not be passed between recruiters or asked to repeat your background to five different people.
Introduction
Forty-five minutes with Tomás. Your background, the shape of the engagement, and your questions about the state of the codebase. No whiteboard.
Technical conversation
Ninety minutes with the CTO. We walk the existing poker specification and the RPS contracts together and talk through how you would attack the shuffle.
Paid test project
A scoped, paid evaluation — a working slice of the real problem. Details on the test-project page. Ten calendar days, and we pay whether or not we proceed.
Review & debrief
You walk two engineers through what you built and defend your design decisions. You receive written feedback in every case, including rejection.
Offer
Commercial terms, a start date, and a handover of the specification, the audit report and the existing contract work.
02/Engagements 02 & 03
The art programme needs designers of its own
These two seats are open for the opposite reason to the poker one. Editions is not paused — it is outgrowing the two people running it, and it has been carried for two quarters by a curator and a CTO making visual decisions that are not theirs to make.
Eighteen collections have shipped and they have sold. They have also been consistent more by luck than by system, and the games estate — fourteen prize tiers and a poker table still to be drawn — has been waiting on someone whose actual job it is.
So we are opening two seats rather than one: a senior designer to art-direct what we release, and a generative artist to build the systems behind the collections that are computed rather than drawn. They are different jobs and we would rather say so than write one advertisement that flatters both.
What you would make
Six surfaces, split across two seats
Between the two roles this is the whole visual output of the company — the collections, the games artwork, what a collector physically receives, and the system that keeps property, art and games looking like one business.
Collection art direction
The through-line of a release: what the work is, how many editions it should be, what the rules of the set are and where it stops. Four collections a year, made with the artist rather than at them.
Generative systems
The code behind the collections that are generated rather than drawn — deterministic from a seed, small enough to store on-chain, and varied in a way that survives someone scrolling all 256 outputs.
The games NFT estate
Fourteen Rock-Paper-Scissors prize tiers from Bronze to Sapphire, where the escalation has to be legible at a glance, plus the table felt, card faces and chip design for poker.
Certificates & provenance artefacts
What a collector actually receives alongside the token: the certificate, the condition report layout, the plot files. The parts of the record a person reads rather than a contract does.
Mint & collection surfaces
The pages a collection is bought from, designed against the same system as the rest of this site. You will work with the Platform squad, and you will win most of the arguments.
The system itself
One visual language across property, art and games. Right now that language is set by the property programme by default, and it needs someone whose actual job it is.
03/Engagement 02
Senior NFT designer — Editions
You decide what a collection looks like. The Director of Editions chooses the artist and sets the curatorial line; everything downstream of that — the set, the system, the type, the certificate, the mint page — is yours.
At a glance
- Engagement
- Staff, or long-term contract
- Shape
- 4 collections a year, plus games
- Location
- Remote, or Austin if you want a desk
- Compensation
- $145k–$185k, or day rate
- Reports to
- Director of Editions
- Start
- Immediate — two seats open
What we genuinely need
- A body of finished, released visual work with a point of view we can actually describe after looking at it. Volume is not the signal; coherence is.
- A systems approach to image-making. A collection is 256 objects that have to read as one work and still reward looking at any single one of them.
- Command of typography and layout — you will own collection pages, certificates and the mint flow, not only the artwork inside the frame.
- An understanding that the image is half the artefact. Token metadata, edition structure and how a work is described in the record are design decisions, and they are yours.
- Comfort working directly with contract engineers, including being told that something beautiful will not fit on-chain and having a second idea ready.
What would help, but is not a filter
- Physical output experience — plotter, risograph, screenprint, letterpress. Half of what we release ends up on paper or canvas.
- Prior NFT collection art direction, and an honest view of what went wrong on it.
- Motion design, for the moving-image collections and the games estate.
- Reading knowledge of a creative-coding environment, even if you do not write it yourself.
Apply
No cover letter. Send the work — a repository, a paper, a portfolio, a plot, a contract you are proud of. Anything that shows how you think.
04/Engagement 03
Generative artist / creative technologist
You write the systems that produce the work. The constraint that makes this role interesting is permanence: the renderer ships on-chain with the token, so the same seed has to produce the same artwork on hardware nobody has built yet.
At a glance
- Engagement
- Per-collection contract, or staff
- Shape
- 2–3 generative collections a year
- Location
- Remote — no timezone requirement
- Compensation
- $160k–$210k, or per-collection fee
- Reports to
- Director of Editions, dotted line to CTO
- Start
- Immediate — two seats open
What we genuinely need
- Production creative coding — p5.js, three.js, raw WebGL and GLSL, openFrameworks, or a plotter toolchain. We care what you can ship, not which library you prefer.
- Determinism as a discipline: the same seed must produce the same artwork on hardware you have never seen, a decade after you wrote it. This is the whole job.
- Judgement about variation. Being able to tell an output set that is genuinely varied from one that is superficially varied is the rarest thing we screen for.
- The ability to reason about on-chain storage limits, and to make a renderer small enough to live in the contract without gutting the work.
- A public body of generative work — releases, plots, shader studies, sketches. We will look at the outputs and we will read the code.
What would help, but is not a filter
- Prior work released on Tezos, or on any platform where the renderer is stored with the token.
- Shader and GPU work, particularly anything real-time.
- Pen-plotter or other physical output from a generative source.
- Colour science, or a considered position on why your palettes work.
Apply
No cover letter. Send the work — a repository, a paper, a portfolio, a plot, a contract you are proud of. Anything that shows how you think.
05/Design process
The same five steps, different middle
Identical in shape to the engineering process, because we think a paid evaluation and written feedback are owed to everyone, not only to engineers.
Introduction
Forty-five minutes with Tomás. Your background, which of the two seats fits you, and your questions about the programme. Send work, not a cover letter.
Work conversation
Ninety minutes with the Director of Editions. We go through three pieces of your work in detail — what the constraints were, what you cut, and what you would do differently now.
Paid commission
A narrow, paid brief with a real deliverable. Ten calendar days, $2,000 on submission, and we pay whether or not we proceed. It is not work we ship.
Review & debrief
You present the commission to the Director of Editions and one contract engineer. You receive written feedback in every case, including rejection.
Offer
Terms, a start date, and the first collection brief — which will already be waiting, because it has been waiting for two quarters.
06/The paid commission
Ten days, $2,000, and we do not ship it
The design equivalent of our engineering test project. It is a narrowed slice of a real problem — the Rock-Paper-Scissors prize ladder — chosen because it exposes systems thinking rather than rendering ability.
A five-tier slice of the prize ladder
Design five of the fourteen Rock-Paper-Scissors tiers — Bronze, Iron, Gold, Osmium and Sapphire. The scarcity has to be readable without the label, and the set has to hold together at 64 pixels and at 2,000.
The rule you worked to
One page describing the system behind the five: what varies, what is fixed, and how tiers six through fourteen would be generated from the same rule without you drawing them.
One token record
The TZIP-21 metadata for a single tier, laid out as you would want a collector to read it. Tell us what you added beyond the required fields and why it belongs in the permanent record.
Generative applicants only — an output set
In place of item 01, a seeded generator producing at least 200 outputs, with the code. We will run it ourselves, twice, and check the two runs match.
Anything you think we have got wrong
Optional and deliberately open. If the brief above is the wrong brief, ignore it and show us the better one — but say why.
Ground rules
- Ten calendar days from the day we send the brief. Ask for more and we will almost always agree.
- $2,000 on submission, whether or not we make an offer. Teams are not applicable here — these are individual seats.
- Any tool you like, including generative and AI tooling, provided you can account for every decision in the review. Disclose what you used; we care about the reasoning, not the purity.
- Deliver as flat files plus a short README, or a repository if there is code. No presentation deck required and none expected.
- The commission stays yours. We do not ship it, and we do not take a licence to it.
What we hand you
- The Editions brand book and the TZIP-21 metadata schema we mint against.
- Every asset from the eighteen collections released so far, including the ones we consider failures, with notes on why.
- The full Rock-Paper-Scissors tier specification — odds, categories and the naming already committed on-chain.
- An hour of the Director of Editions' time during the window, and access to a contract engineer for on-chain storage questions.
07/Questions
The things candidates actually ask
Why can't your own engineers finish the poker project?
They could — that is exactly the problem. The three engineers who know the domain best are committed through Q4 to migrating our settlement layer, which sits underneath every property and every edition we have issued. That work cannot slip and it cannot be handed to someone unfamiliar with our contracts. Poker is genuinely important to us, but it is not the thing that breaks the company if it is late. Rather than starve it of attention for six months, we would rather hand it to someone who can make it their whole focus.
Why are the design roles open at the same time?
For the opposite reason. The art programme is not paused — it is growing faster than the two people running it can art-direct, and it has been carried for two quarters by a curator and a CTO making visual decisions that are not really theirs to make. The collections have shipped and they have sold; they have also been more consistent by luck than by system. Those are two seats we should have opened a year ago.
Is the poker work a rescue of a failing codebase?
No, and we would tell you if it were. There is a detailed specification, an architecture the CTO has already reviewed, and roughly three weeks of exploratory contract work. There is no half-finished product to untangle. You would be starting from a well-lit position, which is precisely why we would rather hand it over now than after a year of intermittent effort.
Individual or agency?
For poker, either — we have scoped the work so that one very strong engineer can complete it in about seven months, or a team of three in roughly half that. Agencies should bid fixed-price against the specification. The two design seats are individual roles: art direction by committee is how a collection stops being one work.
Do I need existing Tezos experience?
It helps, but it is not a filter for any of the three roles. Engineers strong on cryptography and real-time systems who have written contracts on any chain will have their LIGO ramp-up funded. Designers and generative artists who have released on any platform will be fine — what we need you to understand is why the renderer being on-chain changes what you can make, and that takes an afternoon to explain.
Would the design roles work on the art, or the games?
Both, and that is deliberate rather than a workload compromise. The fourteen-tier prize ladder in Rock-Paper-Scissors is a generative edition problem wearing a game's clothes, and the person who solves it well is the same person who should be art-directing a 256-piece collection. If you want to work on only one of the two, say so at the introduction and we will tell you honestly whether that works.
Do you generate the collection artwork with AI?
Not as the work itself. Every collection is authored by a named artist we have met, and the primary split and the royalty go to that person — a generated set with no author to pay would break the only part of this programme we actually care about. Tooling is a different question: artists use whatever is in their studio, and we ask them to disclose it in the token metadata so the record is accurate. The same standard applies to the paid commission.
What does 'full ownership' actually mean?
For poker: you set the architecture, the schedule and the definition of done. The CTO reviews contract code because anything touching escrow gets a second pair of eyes — a standing rule, not a comment on you. For the design seats: you own the visual decision and the Director of Editions owns the curatorial one, which in practice means she chooses the artist and you decide what the collection looks like.
Is the paid evaluation real work you'll ship?
No, for either discipline. Both are deliberately narrowed slices — representative but not useful to us. The poker test project is heads-up only, with no table server, no seats and no side pots. The design commission covers five tiers of a ladder whose final artwork will be made from scratch by whoever takes the seat. We pay for both because your time has value, and we do not use either.
I am an artist, not a designer. Can I still work with you?
Yes, and that is a different door. The Editions programme takes submissions year-round and about one in forty becomes a collection; the terms are published on the art page rather than negotiated in private, and they do not change based on how well known you are. Write to the Editions address with work rather than a proposal.
Apply
Send us something real
No cover letter, for any of the three. Engineers write to the engineering address, designers and generative artists to the design one. Artists who want to be in a collection rather than to design one should write to Editions instead.
Everyone who completes a paid evaluation receives written feedback