Text-heavy lessons.
Children who need reading support get lost in pages of dense text long before they reach the idea behind it.
More time isn’t the answer. Better delivery is.
School curriculum. Made to fit every child.
NeuroLearn is a transformation engine: it takes the curriculum a school already teaches and converts it, through a validated multi-stage pipeline, into formats matched to each child’s neurocognitive profile - attention, reading, working memory, cognitive load and beyond. It changes how a lesson is delivered, not what is taught.
Built for learners with dyslexia, ADHD and other learning differences - and for the many who have no label at all.
The People’s Assembly and the Council of Five Hundred
Athenian democracyThe People’s Assembly (Ecclesia)
The principal governing body of democratic Athens was the People’s Assembly, known as the Ecclesia. It met approximately 40 times a year on the Pnyx hill. All free adult male citizens of Athens over the age of 20 had the right to attend, speak, and take part in decisions.
The Council of Five Hundred (Boule)
The Council of Five Hundred served as the working body of the Assembly. Each year, 500 citizens were selected by lot, with 50 members chosen from each of the ten tribes. The Council prepared matters for discussion, examined them in advance, and submitted proposals, called probouleumata, to the Assembly.
Key terms
Ecclesiathe People’s AssemblyBoulethe Council of Five HundredProbouleumataproposals prepared for the Assembly
Standard contentStructureMeaning & framesGenerationReady lesson
Example from our demo: a textbook page on the People’s Assembly of Athens, adapted for a learner who needs shorter chunks, visual structure and frequent check-ins. Tap the checklist.
One textbook page, three lessons - the version follows the learner:
Extra time helps - a different format helps more
Education systems are built mostly for neurotypical children. When a child struggles, the usual answer is more time, a tutor or more help at home, while the lesson itself stays the same.
Attention that slips away.
Children with attention difficulties lose focus and motivation quickly when a lesson is one long stream.
Girls who go unnoticed.
Girls often mask neurodivergence more than boys, so their needs are identified later, or not at all.
One curriculum. A controlled transformation layer on top.
NeuroLearn is one multi-stage, controlled pipeline that sits on top of the existing curriculum - structuring each lesson, validating it at every stage, and reassembling it into formats built for each learner profile, without altering what is taught.
- A supplement to the school programme, not a replacement - an extra tool for the learner and the tutor
- Works with your existing curriculum
- Adapts to each child’s profile
- Teachers stay in control
- Reproducible, whichever AI model runs underneath
Learner-readyFor every profile
AttentionReadingWorking memorySensory processing
NeuroLearnControlled transformation layer
StructureMeaningFramesGeneration
Your curriculumThe lesson the school already teachesStays exactly as it isAdds to the lesson - never replaces it
Infrastructure, not another app.
Schools, universities and learning centres build their own programmes on top of NeuroLearn. You set the learner profile and the source material; you get learning materials ready to drop into your own systems.
The transformation pipeline
Standard curriculum in. Structured, validated, adapted learning materials out.
- 1
Standard content.
Any curriculum content, in any format.
- 2
Profile & cohort.
The learner cohort’s neurocognitive profile - attention, reading, working memory, sensory processing - defines which adaptation tools apply. Profiles, not diagnoses.
- 3
Controlled transformation.
One multi-stage pipeline - structure, meaning, learning frames, generation - with validation gates between the stages.
- 4
Personalization.
Relevant student interests are layered in, and a final validation checks the result before release.
- 5
Ready to integrate.
Structured, validated, adapted learning materials for your LMS, website or mobile app.
Validation is embedded at every stage - not a single check at the end.
The full sequence
StructureValidationMeaningValidationFrame planning + profile-based adaptation toolsGeneration planningGenerationValidationStudent interestsGenerationFinal validation
Anonymized learning results refine the rules for the next cohort.
One page in. Nine screens out.
The same physics textbook page on distance–time graphs, transformed three times - a Structured, an Engaging and a Guided version. Same facts, same numbers, same questions; a different lesson each time.
Standard content
1 / 9 · Today’s plan
Where each part of the page went · Guided lesson
Nothing was dropped. Each element of the textbook page became one or two screens - reorganised, not rewritten.
- A land-speed recordThe land-speed record of 2015 - about 1400 km/h (389 m/s), faster than sound in air.A record as a comparisonScreen 6The fact comes late, once the words are known: two bars, two numbers, one question - which is larger? The guide character reacts to the answer.
- Speed and distance, paragraph 1What speed tells you; the units m/s, km/h and mph.Key words, one picture eachScreen 2Speed, unit, m/s, km/h, mph - each on its own line with an icon. The main idea is one sentence: a bigger speed means faster movement.
- Speed and distance, paragraph 2Average speed vs instantaneous speed - two bold terms inside one paragraph.One road, two signsScreen 3The same car at SLOW and FAST along one journey; average and instantaneous become two places on a picture. Self-check: what does a speedometer show?
- Equations + the speed triangleTwo equations, the speed triangle and a “cover the quantity you want” instruction.The triangle, once and fullyScreen 5d on top, v and t below, all three equations listed. The learner meets the tool once, completely - no revisit, no rearranging game.
- A guide character on every screen
- One idea, one picture, one question
- Yes / Pause - a choice, not a demand
- Sound, motion, brightness - under control
A guide character on every screenThe same girl and her dog lead the lesson. A familiar face lowers the cost of every new screen.
One idea, one picture, one questionEvery screen has a single main idea box and a two-option check. Never more than one thing to decide.
Yes / Pause - a choice, not a demandScreen 1 asks “Ready?” with a Pause button. The learner can stop or go back at any time, and is told so.
Sound, motion, brightness - under controlSensory toggles sit on every screen, plus “How do you feel?” so the teacher sees energy dropping before it does.
Every number, term and question from the page is still there. What is taught didn’t change - how it is delivered did.
Show the full case studyAI accelerates. Validation and experts ensure quality.
NeuroLearn isn’t a text rewriter - it’s a controlled transformation system. It draws on research and established pedagogical approaches - Orton-Gillingham, Structured Literacy, UDL, TEACCH®-informed principles and emotional regulation frameworks - and translates them into rules, algorithms, adaptation mechanisms and validation methods, not prompts. The AI models inside are interchangeable; the methodology stays ours.
Built on established educational approachesPrinciples, not licensed programmes - translated into NeuroLearn’s rules and validation
- Orton-Gillingham–informed structured literacyPrinciples of explicit, structured, sequential and multisensory literacy instruction.
- Structured Literacy informed by Wilson Reading System®Relevant principles of systematic literacy instruction; the transformation and validation mechanisms are developed within NeuroLearn.
- Structured learning informed by the TEACCH® approachStructured environments, visual supports, individualised organisation and predictable learning routines.
- Universal Design for Learning (UDL)Multiple ways for learners to engage with content, understand information and show what they know.
- Emotional and behavioural regulation principlesAdapting learning experiences to learners’ emotional and self-regulation needs.
Controlled by design
A multi-layer validation system: AI works inside programmatically bound constraints, validation gates sit between the stages of the pipeline, and expert educators validate the rules upstream - before anything reaches a classroom.
Deterministic and reproducible.
Expert decisions at key points are built into deterministic rules, so the same lesson for the same learner profile gives a reproducible result and teaching methods are followed precisely.
Frames, not free generation.
AI works inside validated structures and pedagogical rules. It changes how content is delivered, never what is taught.
Experts in the loop - upstream.
Educators and methodologists shape and validate the rules for each profile and cohort. The classroom teacher doesn’t have to review every adaptation.
Profiles, not diagnoses.
A learner profile describes which adaptations to apply. We don’t store medical diagnoses.
Small, safe personalization.
Real-time adjustments are limited to things like the child’s interests and pace.
Our core asset
The validation methodology is NeuroLearn’s key technological asset. The LLMs that parse and generate content are a replaceable layer.
Technology resilience.
No lock-in to one model generation or vendor - as the field moves, we adopt what is better and the core stays where it is.
Flexible cost management.
Every stage of the pipeline can run on the model with the best balance of quality, speed and cost.
The people behind NeuroLearn
Pedagogy, clinical psychiatry, product and engineering in one team - because a lesson for a neurodivergent child has to be right on all four counts at once.
- Founder
Maja Erčulj
CEO · FounderMaja Erčulj founded NeuroLearn and leads its partnerships and fundraising across the EU. She established and runs NeuroLearn, zavod - the non-profit that works with EU schools and produces the learning materials.
Andrei Sobolev
CPOAndrei Sobolev has worked in digital health since 2012. He founded and led a telemedicine platform supporting post-surgical and chronic-care patients in a strictly regulated medical domain, and took it through a successful exit. At NeuroLearn he brings that experience of sensitive, regulated domains to product strategy and operations.
Karine Petrosyan
Head of Pedagogy & Learning MethodologyKarine Petrosyan has spent more than 25 years as an educator working with neurodivergent children. She is dyslexic herself - which is not a footnote to her work, but where it started. Before NeuroLearn, she worked with NGOs drafting gender equality legislation, and has spent her life championing the potential of those the system was never built for.
Dmitrii Popov
CTODmitrii Popov is a senior technical leader with more than 15 years of building and scaling complex systems across telemedicine, fintech, agriculture and consumer products. He specialises in backend architecture and in systems that must evolve as requirements change, working comfortably between R&D and production. At NeuroLearn he designs the platform’s architecture and leads its development.
Alicija Moroz
Psychiatrist & PsychotherapistAlicija Moroz is a psychiatrist and psychotherapist specialising in neurodevelopment and cognitive functioning. At NeuroLearn she leads the work on learners’ individual neurocognitive profiles - shaping how learning material adapts to attention, memory, executive functions, sensory processing and cognitive flexibility. Her focus covers both learning difficulties and the high cognitive potential that calls for more challenging material.
Let’s shape personalized learning together
Who we’re looking for: schools · Learning centres · Educational institutions · Researchers and experts
Prefer to talk first?WhatsApp us - usually answered the same day
One class, one subject, one term. Your lessons stay as they are - we give the children who need it an adapted supplement before or after class, and together we measure what changed.