About Artificial Societies
Artificial Societies (societies.ai) helps leading organisations understand how real-world audiences think, feel, and respond without the cost, delay, or limits of traditional research. We build large-scale simulated populations grounded in real-world data, so clients can test messages, strategies, and concepts and get insights backed by millions of responses in hours rather than months, including from audiences research normally can't reach: investors, CEOs, opinion leaders.
We're backed by Y Combinator, Point72 Ventures, and Kindred Capital, plus investors from DeepMind and Sequoia Scout. We're a lean team that moves fast, holds a high bar, and trusts each other to own outcomes end-to-end. Ideas win on merit, not seniority.
The Role
Most people have never seen a simulated society at work. The fastest way to make them understand it is to put one in front of them. Some things we want to build, to give you the register:
- A simulation of the Roman Senate, with a live Fox News feed piped in as the day's events. What would Cato make of the news cycle?
- Simulated prediction markets. Take live Polymarket topics, run them through different scenarios, and watch what shapes the opinions that shape the world.
- Whatever you come up with that's better than both.
You'll work directly with James, our CEO. You'll live in Claude Code and Python notebooks, find datasets nobody has thought to use, wire them into our simulation engine, and ship a front end that stands on its own. Then you'll make sure people actually see it.
What You'll Do
- Invent and ship public demos. Own the whole arc: the idea, the data, the simulation design, the front end, the launch.
- Find novel datasets. Work out what data exists in the world, how to get it, and how to turn it into simulations.
- Make the front end good. These demos are how people form their first impression of simulation. They need to look like something worth sharing.
- Get attention. Understand what travels on X, LinkedIn, Reddit, and Hacker News, and build with distribution in mind from the first line of code.
- Feed it back into the product. Building at the edge of what our engine can do surfaces what it should do next. You'll take that straight to our research and product engineers.
What We're Looking For
- You've shipped things end-to-end. Data, backend, front end, deploy, launch. You have a list of things you've put on the internet, and at least one that people paid attention to. You can explain why it worked.
- Fluent with AI coding tools. You use Claude Code or equivalent daily, and you know where it stops being trustworthy.
- Comfortable in Python notebooks. You can pull a messy dataset apart, reason about what it actually says, and be honest when it doesn't say what you hoped.
- Front-end capable. You can build an interface people want to click on. Design taste matters more than framework mastery.
- A source of ideas. You'll be judged partly on the quality of what you bring, not just your execution of ours. We want the person whose Slack messages start with "okay this is insane but". You don't wait for a ticket. You see something worth building, and you build it.
Nice to Have
- Ex-founder. A huge plus. If you've run your own thing, funded or not, alive or dead, you already know how this role works.
- An audience of your own, or a track record of things that gathered a following.
- Interest in politics, forecasting, history, or media: any domain where social opinion is the thing that matters.
- Familiarity with TypeScript (our product stack), or a background in social science, network analysis, or data journalism.
Logistics
- Location: London, in-office. We build better together.
- Salary: £70,000 to £100,000 depending on experience, plus equity.
Why Join Now
Simulation is a brand new technology, and most of the world can't imagine what it can do yet. This is the role that changes that. You'll have unusual latitude: direct access to the CEO, a frontier simulation engine to point at whatever you find interesting, and permission to build the ambitious thing rather than the safe one.
How to Apply
Send us the three most interesting things you've built and put in front of people. Then tell us, in a paragraph, the simulation you'd build in your first month and who you'd want to see it.