Software Engineer

We’re an early-stage team building our way toward the answers. We form hypotheses, make things, test them, learn, and sometimes tear them down and start again.

We are looking for a Director of Product Management to help turn those ideas into a product that actually works.

Who are you?

You are a product-minded engineer who has learned to build with AI without confusing generated code with good engineering.

You know the craft deeply: fundamentals, system design, and what it takes to operate something once real people depend on it. You hold a high bar for correctness, security, and maintainability, and you don’t relax it because a model wrote the first draft.

But you’ve also changed how you work. You use tools like Codex and Claude Code across planning, implementation, testing, debugging, review, and documentation, and you move unusually fast because of it. You spend less time writing code and more time judging it: reading, testing, rejecting, and deciding what deserves to exist. You instinctively notice repetitive work and ask whether it should exist at all.

You understand that building AI products is not the same as building traditional software. Prompts, models, context, evaluations, agent behavior, traces, failure analysis, and human review are part of the development lifecycle, not implementation details that come after the engineering work is done. Probabilistic behavior is a design problem you take seriously rather than a defect you expect to eliminate.

You also understand that, here, learning is product and product is learning. You don’t need to be the learning expert, but you will be building knowledge graphs, learner state, agent behavior, and assessment machinery, and you need to engage seriously with what those things are for rather than treating them as requirements handed to you by someone else.

You like hard, underspecified problems. You can turn an early idea into something working, put it in front of learners, and throw it away when the evidence says so.

What will you do?

As a Senior Software Engineer, you will design, build, and operate the platform: the learner-facing experience, the services behind it, and the AI systems that do the teaching.

You will work across the stack, in close partnership with the Director of Learning and the Director of Product Management.

With Learning, you will build the machinery behind the learning model: competencies, knowledge graphs, learner state, agent behavior, and assessment. Learning decides what good teaching looks like. You decide how it is represented and how it runs. Those two decisions constrain each other, so you will make them together rather than in sequence.

With Product Management, you will turn strategy into a roadmap that can actually be built: scoping, sequencing, tradeoffs, and the AI development lifecycle of prompts, evaluations, traces, and failure analysis. You are expected to shape what gets built, not simply receive it.

Both leads are hands-on and prototype in code, so you will often be working from a half-built prototype as much as from a written specification. You will also work with Experience Design, and with the educators and subject-matter experts who build course content. This is an early-stage team, so you will help shape how we engineer as well: our practices, our tooling, and our standards.

Responsibilities

• Build the product end to end. Design, build, and ship learner-facing experiences and the platform services behind them, and stay responsible for them once they are running.
• Build how AI teaches. Develop the AI-powered teaching, assessment, feedback, and content-authoring capabilities at the center of the product, working with Learning on agent behavior, prompts, and constraints.
• Make AI behavior measurable. Build the evaluations, observability, tracing, and safeguards that tell us whether the system is actually working, and catch it when it isn’t.
• Prototype, test, and productionize. Turn early ideas into working prototypes quickly, build the testing Product Management and Learning need to run against them with synthetic and real learners, and evolve the ones that survive into reliable production systems.
• Engineer with AI, at a high bar. Use AI development tools throughout planning, implementation, testing, debugging, review, and documentation, and hold generated work to the same standard as anything you would write yourself.
• Own the architecture you build on. Make sound tradeoffs about data models, services, and system design, and treat knowledge graphs, learner state, and event flows as things you design deliberately rather than accrete.
• Help build the engineering practice. Establish lightweight approaches to review, testing, deployment, and operations that create rigor without creating bureaucracy.

This is a senior, hands-on role. We expect you to bring substantial engineering experience and enough depth in the craft to make consequential architectural decisions and establish strong practices in an early-stage company. We care much more about what you have actually built and how you operate than about a particular degree, title, or number of years.

• Strong engineering fundamentals. Data structures, algorithms, system design, and software architecture. Applied, not recited.
• Experience building and operating real systems. You have shipped high-quality web applications or distributed systems and kept them running once they mattered.
• Substantial hands-on experience building generative-AI applications. Agents, tool use, retrieval, structured generation, context management, and AI evaluations. Built, not just read about.
• A high bar for AI-assisted work. You can evaluate generated code, debug difficult problems, and tell the difference between code that runs and code that is right.
• Product judgment. You can turn early, sometimes ambiguous ideas into working products, work from detailed specifications, and push back when one is solving the wrong problem.
• Systems thinking. You naturally see dependencies and second-order effects across experience, data, learning, AI behavior, and operations.
• Exceptional resourcefulness. You learn new domains quickly, operate well without a large team or established process, and can get excited about lighting yesterday’s great design on fire when you find a better one.

Helpful, but not required

• Experience with TypeScript, Node.js, Svelte, or another modern web framework.
• Experience with PostgreSQL, vector search, knowledge graphs, or event-driven systems.
• Familiarity with learning platforms, competency-based education, assessment, accessibility, or multilingual products

We’re a product company building the underlying platform for a new model of AI-native education-beginning with our first application: an AI-native university in the UK.

We believe AI gives us the opportunity-and obligation-to rethink some of the basic architecture of education: what people should learn, how teaching works, what assessment means when AI is ubiquitous, how learning adapts to an individual, and where humans matter most.