Semilattice / Careers / Founding Full-Stack Software Engineer

Help us simulate the human world

We are obsessed with complexity. Specifically: the emergent complexity of human systems like organisations, markets, communities, and countries, and the uncertainty that complexity generates. We believe that uncertainty is a core reason why building things is so hard: 90% of startups fail, 95% of new products fail, 80% of government change efforts fail, and 90% of infrastructure projects go over budget. We believe far better ways of managing and governing exist in the search space, but we need new computational tools to carry out the search.

Historically, human systems have been beyond the modelling capabilities of both the brain and computers due to the individual-level complexity of humans. You can build a digital twin of any machine but as soon as you add a human its predictive accuracy breaks down. With ~all of the internet’s human behavioural data in their parameter space, LLMs change this. Along with other machine learning and data science techniques, and with the right implementation, we are betting that LLMs unlock new kinds of computational models of human systems, enabling decision-makers to accurately simulate counterfactuals and make better decisions.

Our mission is to simulate the human world to advance society, and this year we shipped our first step. Semilattice Answers accurately simulates how groups of people answer questions, targeting the ~$100B research market. Some of the problems we’re now thinking about include: how to fit the product to decision-makers’ workflows and make it grow; how to real-time align audience models with shifting public opinion; how to model the information diets for specific micro populations; how to locate, ingest, and apply the large amounts of human system data we’ll need; and how to expand the capabilities of the underlying simulation engine to model multimodal, interactive, and temporally deep behaviour.

Join the founding team

We have just raised a pre-seed financing round from a set of leading UK and US investors. Joining now means signing up to the hard and messy challenge of finding product-market fit. If we can pull it off, the company will be on a growth trajectory which will mean rapid career progression and abundant learning opportunities.

Semilattice is founded by Joseph Wright (Ex-Google, Stripe, Figma) and Fabian Teichmueller (ex-Google, Kaluza) with over 30 years combined product, engineer, marketing, and commercial experience. We are building the company from the ground up for the long term and are committed to finding the right first employees to set the culture. You can learn more about our mission, vision, and technology thesis in our founders’ letter:

Humans + Time

About the role

We have an LLM-based audience simulation model in production (Python) benchmarked at >87% accuracy and a v1 app built on Next.js (TypeScript) and Supabase (Postgres) for decision-makers to create models of their audiences and ask questions.

Pre-product-market fit, our roadmap is a set of branches we’ll iteratively explore as we learn. We are looking for a Founding Full-Stack Software Engineer to:

• Design and implement improvements to our existing LLM-based answers simulation engine, expanding its use cases and tuning its alignment with reality.
• Talk to users to understand their use cases and workflows and collaborate with the founders on decoding and navigating the market to decide what to build.
• Build user-focused product features to increase our existing product’s market-fit and iteratively improve, adapt, or kill those features based on further feedback.
• Research and develop technical solutions for new and unforeseen problems to enable the company to rapidly explore the PMF search space.
• Identify, reason about, and process real-world human behavioural data to train, prompt, and evaluate human simulation models.
• Design, implement, and maintain our infrastructure for model evaluation, deployment, and testing.
• Own a range of ongoing engineering decisions, from high-level architecture and technology questions to low-level implementation details.
• Help create the Semilattice Engineering culture, developing processes, rituals, and tooling.

About you

• You have a range of engineering experience on production software across both front-end and back-end, giving you the knowledge and confidence to make high and low-level engineering decisions.
• You are proficient in, or able to quickly become proficient in, TypeScript and Python.
• You have experience with or are willing to learn our current tech stack: Next.js (TypeScript), Supabase (Postgres), OpenAI/other LLM APIs.
• You are excited to join, and ideally have experience of, a very early-stage, pre-PMF startup: tiny team, early technology, lots of uncertainty and urgency.
• You are flexible and scrappy, willing to routinely take the initiative to find novel solutions to unexpected problems.
• You are aligned to the mission to simulate the human world to advance society. Ideally, you have worked on or tinkered with simulation or AI-based approaches to modelling and predicting human behaviour.
• You are willing to work in-person at least 3 days a week from central London.

Benefits & compensation

• Stock option plan with above-market equity.
• Salary of £70k to 100k depending on experience.
• 30 days paid time off.
• Budget to buy the equipment you need to be productive.
• Central London office.
• 10% “System Time”: one protected, self-managed day every 2 weeks to work on improving the system of the company and/or experimental ideas.
• Plus the standard benefits:
  • Parental leave.
  • Pension.
  • Public holidays.
  • Sick days.

Apply

To apply, please send your resume to [email protected]