Full Stack Engineer

Lexplore uses AI and eye-tracking technology to measure reading ability in school-age children. Thousands of schools across the Nordics use it to find the children who need support early, and the children who use it read measurably better. We are now changing what the product is, and we are looking for an engineer to help decide what it becomes. The short version Role: Full Stack EngineerExperience: 3-7 years building and running production systemsCore stack: .NET, Azure, React/TypeScript, Python and AI/MLLocation: Stockholm. Hybrid, with set days in the office each weekTeam: The first engineer joining the team we are building now. You would report to our CPTOLanguage: Swedish and/or EnglishWork permit: We are not able to sponsor work permits or relocation, so you need an existing right to work in SwedenEmployment: Permanent, full-timeStart: As soon as possible About Lexplore A child reads a short text on screen, we track their eye movements while they do it, and our AI models turn that into an objective picture of their reading development. It takes a few minutes, and any teacher or parent can run it without training. Behind that sits three decades of research and a normed dataset built from more than a million assessments. That dataset is what our machine learning models are trained and normed against, and it is what makes the result trustworthy rather than just fast. What is changing is the shape of it. We are moving from a standalone tool towards an intelligence layer that lives inside other platforms and services, while building an offering directly to families. What you would actually work on Two growth tracks, and the work that makes both of them possible. Not all of it at once, and not all of it alone. Growth track 1: Integrations Integrations are how Lexplore reaches more readers. A working integration layer and a live demo are already in place, and now we are building it out at scale. We are expanding the API and backend that lets LMS platforms, literacy tools, and educational publishers launch a Lexplore assessment directly from their own product and receive the results wherever they need them. Growth track 2: Direct to families Every parent wants to know how their child is doing at reading. Today that mostly depends on whether their school happens to use Lexplore, which makes reaching families directly a substantial opportunity for us. We are building the version a parent buys directly, for every child rather than only those already struggling. It is an early-stage consumer product inside an established company, and you would help shape what it becomes, how it scales, and how it earns a parent's trust. Making both possible: Keep it solid Strengthening CI/CD, monitoring and access managementMaking architectural knowledge shared across the team as it growsPaying down the technical debt that gets in the way, and leaving the rest aloneFinishing the move off the Windows client. The web assessment (Flutter Web) is the current product and the .NET WPF client is being phased out, and completing that removes a large maintenance burden The engineering The models that turn raw gaze into a reading profile are the core of the product, and improving them is ongoing work rather than a finished job.Eye-tracking produces high-frequency time-series data. Turning that into a reliable pedagogical signal is a real data engineering and ML problem, not a dashboard.We handle sensitive personal data about children at scale, under GDPR and across several national school systems with different rules, identity providers and expectations.Integration work here means becoming part of the daily infrastructure of a school. The reliability bar is set by a teacher with thirty children in front of them. The platform Frontend: React / TypeScript (teacher portal)Backend: .NET APIsAssessment, web: Flutter Web (current product)Assessment, Windows: .NET WPF (legacy, being phased out)Data pipeline: C# for event extraction, Python for feature extraction and ML modelsInfrastructure: Azure, Blob Storage, Service BusDatabases: SQL Elastic Pool, RedisAuth: Azure AD B2C, SAMLIntegrations: Skolon, Wonde, Feide, and more coming What we need from you We care more about how you work than whether you tick every box. If you are strong on the first list and curious about the second, apply. Solid from day one 3-7 years of professional experience, including .NET backend engineering (Java or Node.js also relevant) in productionAzure cloud environmentsDevOps and CI/CDComfort navigating a complex system you did not build Useful, or something you can grow into Consumer or B2C product experience, particularly self-serve signup and onboardingIntegration and API design patternsReact / TypeScriptPython and data engineering The person we are hoping to find Hands-on: You would rather build it than delegate itPragmatic: You find the right balance between speed and quality, and you know the answer changes by contextCommunicative: You share what you know, and you ask when you do notCurious: You pick up new technology because you enjoy itComfortable with change: A company changing direction reads to you as an opportunity, not a warning sign Who you would be working with You would work most closely with our CPTO, but we are small enough that you would work with everyone: product, the people talking to schools, whoever happens to be closest to the problem that week. High pace, short decision paths, and nobody standing between you and the decision. You would be the first engineer joining the team we are building now, which gives you a large say in the tooling and the practices that team grows into. The platform is not a blank page: it is a live product that engineers before you built and that thousands of schools depend on, so part of the job early on is learning it properly. What we offer Work that measurably helps children learn to readA large say in the tooling, standards and ways of working the new team grows intoReal influence over technical direction, and a say in what gets built as well as howHybrid working from our Stockholm office, with set days on site so the team is there at the same time The process Introductory call (30 minutes)Interview with our CPTOTechnical interviewInterview with our CEOReferences We review continuously and close the ad once we have found the right person, so apply sooner rather than later.