An offline-first learning environment where students use executable pseudocode to understand algorithms, data, AI, systems and software, without getting stuck on programming syntax.
Modern AI tools can generate working software from natural-language instructions. The role of manually memorising programming syntax is changing rapidly. Computer science itself is not.
Generative AI can turn natural-language instructions into working software. That makes memorising programming syntax a less valuable starting point for beginners.
Students still need to understand algorithms, data, logic, systems and how software behaves, to direct AI tools, evaluate outputs, and solve the real problem underneath.
Thadus removes the syntax burden so students can focus on those concepts directly, before they ever worry about how a particular language requires them to be written.
AI can write the code. Thadus teaches students how to understand it.
Travelling across Asia, through classrooms, community centres, and conversations with young people who were sharp, motivated and completely locked out of computer science, we kept seeing the same pattern. The barrier wasn't aptitude. It was access: no reliable internet, underpowered devices, and learning environments that required a student to master complicated programming syntax before they could explore a single concept.
That created a clear problem to solve. In 2025 we founded Thadus Group and built Thadus CodeLabs: an offline-first computer science learning environment built around one central idea: beginners should be able to understand how software works before they have to master how to write it.
Thadus CodeLabs is a desktop learning environment built around one idea: beginners should be able to explore computer science before they have to master a programming language. Students open it and start reasoning computationally from day one.
Twelve structured courses take students from computational thinking through algorithms, data, games, applications, cybersecurity, AI and machine learning foundations, all within one consistent environment, with no syntax getting in the way of the ideas.
// Create and move an object ball = new object ball moveRight(15) ball moveUp(10) // A simple condition num score = 8 if (score > 6) print levelclear else print retry end if
Students write almost plain-English instructions and immediately see how their logic affects the system. The objective isn't to make programming easier. It's to make computer science understandable.
Thadus deliberately minimises programming syntax so students can concentrate on what a program is doing, rather than how a particular language requires it to be written.
Students focus on algorithms, logic, data, systems and computational thinking, not on the bookkeeping rules of a particular programming language.
Instructions are written in simple, readable commands that behave like real programs. Students can express a process almost like writing it in plain English, and see it actually execute.
Students aren't memorising Thadus syntax for its own sake. They are building mental models they can apply to any programming language, or AI-assisted development tool, from a position of understanding.
Most educational platforms solve one access problem. Thadus solves two simultaneously.
Runs entirely offline on low-spec hardware. No internet, no accounts, no installation headaches. Computer science education becomes possible where connectivity cannot be assumed.
Uses executable pseudocode so students focus on understanding concepts instead of fighting programming syntax. The cognitive barrier comes down alongside the digital one.
A beginner trying to learn programming is often asked to understand two things simultaneously: what the computer is supposed to do, and the rules of a particular programming language. Thadus separates those two problems.
Students first learn to reason computationally: to express processes, understand systems, and think in algorithms. The syntax can come later. Once students understand the underlying concepts, they can learn any programming language, or work with AI coding tools, from a position of understanding rather than memorisation.
The skills behind technology are changing.
AI and big data, technological literacy and systems thinking are among the fastest-growing skill areas identified by employers globally through 2030.
Programming languages change. Frameworks change. AI tools change. The underlying concepts don't disappear. Every course in Thadus CodeLabs is built around ideas that remain relevant regardless of which language or platform eventually produces the code.
Students explore classification, recommendation, labelled data, prediction and AI decision logic, building an intuitive understanding of how machine-learning systems work, without the black box.
Students learn to collect, transform, interpret and visualise information, building charts, detecting outliers and making evidence-based decisions from real datasets.
Students learn how complex problems can be broken into repeatable, logical processes: sequencing, searching, sorting and decomposition as foundational thinking tools.
Students explore authentication, permissions, access control and suspicious behaviour, understanding how secure systems are designed to protect and verify.
Students learn state, events, conditions, iteration and interactions, the same patterns that underpin real game engines, simulations and reactive interfaces.
Students learn interfaces, components, navigation and user flows, building mental models they can transfer directly to any app development environment.
Tools change. Understanding compounds.
The curriculum doesn't start with "learn a programming language." It starts with how computers think, how systems work, and how technology solves problems. Twelve courses. Two per month. Built on CSTA K-12 Computer Science Standards for Years 5–9.
Sequencing, decomposition, loops, conditions, and functions through logic-based route challenges.
Arrays, search techniques, and foundational sorting logic.
Collecting, transforming, and visualising data: totals, averages, trends, and outliers.
Scene design, sprites, controls, and win/lose logic, ending in a playable mini game.
Interface structure, components, and screen linking in a working mini app.
Multi-screen navigation, onboarding flows, menus, and settings.
Entities, metrics, and rules to model real-world systems, trade-offs and bottlenecks.
Dashboards, alerts, and charts for data-informed decisions.
Password logic, access control, and permissions in secure portal design.
Students build rule-based systems and explore how machines classify inputs, make decisions and generate responses. No black box: students learn what is happening underneath.
Students explore features, labelled data, classification and prediction quality, building an intuitive understanding of how machine-learning systems work before they ever see a framework, from the inside out.
Full product planning and build, culminating in a final showcase project.
Underneath every Thadus course are the same foundational ideas, the concepts students carry forward regardless of which language or tool they eventually use.
Algorithms
Data
Logic
Systems
Problem-solving
Computational thinking
AI concepts
Machine learning
These are the concepts students carry forward, regardless of which language or tool they eventually use.
Twelve courses. Twelve ways to understand technology. Each one builds a concept students can use, regardless of which language or tool they eventually work with.
Students work through route-planning and logic challenges that build precise algorithmic thinking. By the end they can write loops, conditions and reusable functions, the patterns that appear in every type of software.
Students build arrays from real data tables, then write algorithms to search through and sort them. The focus is on understanding why an algorithm works, not just that it does. A foundation for data, AI and systems work in later courses.
Students load real datasets, calculate totals and averages, plot trends and identify outliers. They practise asking "what does the data actually say?", the core skill behind data analysis, business intelligence and machine learning alike.
Students plan, build and publish a playable mini game. Game development is a powerful context for learning conditions, iteration and state, because students immediately see how their logic changes what happens on screen.
Students design and build a multi-screen mini app, starting with a login screen and linking it to a home screen. The focus is on understanding how interface structure and user flow are designed before they are coded. Directly transferable to any app development environment.
Students build a complete multi-screen product with onboarding, navigation and settings. They learn to think from the user's perspective: why a flow works, where it breaks, and how to improve it. This is product thinking, not just coding.
Students model real-world systems: queues, traffic cycles and populations. They observe how rules and variables interact to produce emergent behaviour. Small rule changes, large outcomes. This is the course where systems thinking becomes concrete.
Students build decision dashboards from scratch: loading datasets, plotting attendance and performance trends, detecting outliers and creating alerts that fire when a threshold is crossed. The full pipeline from raw data to a decision.
Students build a secure school portal, writing password rules, adding role-based access permissions, detecting suspicious input and debugging insecure code. Security is a great lens for understanding how systems are designed and where they break down.
Students build an AI assistant from scratch using rules and classifiers, then test it, find its biases, and improve it. No black box. Students see exactly why the system responds the way it does, which means they understand what "AI" is actually doing at the logic level.
Students label datasets, choose features, build a simple classifier and then test it on data the model has never seen. The course is deliberately non-framework: students understand what a model is doing before they ever see one pre-built. That's the knowledge that transfers to any ML context.
Students choose a track (App, Game, Simulation, Dashboard, Cyber, or AI Lab) and build a complete product end to end. This is the course where everything comes together: product thinking, iteration, testing, and delivering something a real person could use.
Offline. Accessible. Concept-first.
