IRL PLAYground
Adaptive Infrastructure for Learning and Play
Role: Founder, Systems Architect & Product Design Lead
Type: Applied AI Platform and Product–Infrastructure Development
Regions: Global (Pilot deployments in the UK and Southeast Asia; debut at the Hong Kong Toy Fair 2025)
Introduction
IRL PLAYground is a study in education and play as interconnected infrastructures, and AI as civic intelligence.
It reimagines the classroom and the toy as parts of a single adaptive system — one that learns from itself through the continuous feedback of people, data, and materials.
Rather than treating software, toys, and educational content as separate products, IRL PLAYground unites them into a cyber-physical ecosystem: objects act as data interfaces, classrooms operate as learning networks, and AI coordinates complexity quietly in the background.
The result is an environment that rebalances in real time, enabling teachers, students, distributors, and manufacturers to co-produce value across a shared infrastructure.
The classroom is not a workflow, and the toy is not an object — both are protocols for learning.
World-Systemic Frame
Education and play are two halves of the same world-system.
Both shape how societies learn, produce, and reproduce meaning — one through governance, the other through material imagination.
IRL PLAYground positions itself at their intersection as a cybernetic infrastructure that connects pedagogy, manufacturing, and distribution into a unified ecology of adaptation.
Within this frame, the classroom becomes the institutional interface, and the toy becomes the material interface of the same system.
Together they form a recursive loop between knowledge, logistics, and ethics — a planetary circuit where data, design, and care circulate continuously.
Learning is not linear. Production is not external.
Both are feedback systems within the same ecology.
Civil / Civic Reciprocity
The IRL PLAYground ecosystem is built around a dual architecture of civil and civic engineering:
- Civil engineering shapes the technical substrate — the software, data models, automation pipelines, and 3D scanning systems that maintain daily operation.
- Civic engineering shapes governance — the participatory frameworks that define how intelligence is trained, shared, and audited across schools, distributors, and designers.
Each informs the other.
Teachers retrain models through use; distributors feed metadata that enriches learning contexts; policymakers set boundaries that classroom feedback then refines.
The system becomes a mutual choreography of precision and empathy — a civic infrastructure where AI learns to care by learning from those it serves.
Complexity as Method
IRL PLAYground is designed not to simplify education or manufacturing, but to sustain their shared complexity.
It models both as feedback systems in which intelligence is distributed across human, digital, and material agents.
- Adaptive learning loops: AI assists teachers and learners in real time — identifying gaps, proposing interventions, and evolving through human correction without enforcing standardisation.
- Ethical loops: educators and partners train small, interpretable AI models aligned with local values and governance frameworks, ensuring transparency, consent, and data sovereignty.
- Material loops: toys are scanned through photogrammetry and structured-light systems, each embedded with metadata defining pedagogical purpose, ecological origin, and supply-chain story.
- Institutional loops: open APIs and transparent protocols connect classrooms, factories, distributors, and national repositories into a single, traceable network.
Implementation
At its technical core, IRL PLAYground is a modular platform combining software, AI, and physical design.
It integrates school data systems, distribution networks, and object-intelligence mapping into a single open framework:
- Software stack: TypeScript, React, Node.js, Python (FastAPI) for adaptive interfaces and automation.
- AI layer: Fine-tuned GPT APIs, TensorFlow Lite, CLIP-based recognition for localised, teacher-led model training.
- 3D & hardware systems: RealityCapture, Agisoft Metashape, OpenMVG, Raspberry Pi sensors, and volumetric cameras for photogrammetry and real-time capture.
- Data architecture: GraphQL schema linking material origin, pedagogical context, and institutional metadata for interoperability.
- Integration: Distributor APIs, SIS connectors, and photogrammetry ingestion pipelines.
Results
- Deployed a real-time adaptive education framework, connecting physical play with digital cognition.
- Reduced administrative overhead by 40% in pilot schools, allowing educators to focus on care and pedagogy.
- Demonstrated AI interoperability across software, hardware, and photogrammetry pipelines — validating the link between material intelligence and digital learning systems.
- Engaged manufacturers, educators, and distributors across Asia and Europe in the PLAYground Partner Programme, co-developing ethical, circular pipelines for AI-linked play environments.
- Established a governance model for educational AI — including role-based permissions, data-visibility frameworks, and teacher-led retraining systems.
Featured Modules
- KIN: Explores modularity and systems thinking through reconfigurable structures that teach algorithmic reasoning via physical assembly.
- Pieces: Provides alternative characters to the animal kingdom, creating a puzzle from abstract sets for AI storytelling.
- Our Home: Integrates puzzle solving and environmental design into first-principles problem solving, bridging the emotional and infrastructural dimensions of ecological thinking in play.
- Produce Crayons: Transforms waste materials into creative tools, connecting ecological cycles with the material economy of education.
Each module operates as both a toy and a data node — scanned, digitised, and networked to form a living archive of learning objects inside the PLAYground ecosystem.
Stakeholders & Coordination
Co-developed by IRL’s design, software, and pedagogy teams in partnership with schools, distributors, and regional education authorities.
The project demonstrates cross-sector interoperability — uniting manufacturing data, AI learning systems, and institutional governance into a coherent infrastructure for adaptive education and play.
Tools & Methods
Software Stack: TypeScript, React, Node.js, Python (FastAPI), PostgreSQL
AI Integration: Fine-tuned GPT APIs, TensorFlow Lite, CLIP-based object recognition, teacher-led retraining tools
3D Systems: RealityCapture, Agisoft Metashape, OpenMVG
Hardware Integration: Raspberry Pi sensors, volumetric cameras, interactive touch grids
Data Infrastructure: GraphQL schema linking material origin, pedagogical context, and usage data
Design Systems: Figma, Tailwind, IRL’s classroom feedback simulators
Civic–Scientific Insight
IRL PLAYground reframes both education and play as infrastructure problems — spaces where care, data, and design co-produce intelligence.
By connecting material and digital systems, it models a civil–civic form of AI: one that learns from the social and ecological patterns it inhabits.
In this vision, each classroom, toy, and networked interaction contributes to a planetary feedback system — transforming global distribution into a medium for collective learning, ethical production, and civic renewal.
Intelligence should not manage the world.
It should help the world learn how to care.