NeuroStell
NeuroStell
Adaptive learning for institutions

Learning systems should adapt to cognition, not the other way around.

NeuroStell is a science-backed adaptive learning platform that renders the same content differently based on attention, working memory, sensory load, and pacing differences. This proof of concept shows how cognitive profiling, adaptive rendering, and multisensory delivery can work together without diagnosing anyone.

Interpretability first
Rule + ML hybrid
LMS ready (SCORM/xAPI)
GDPR aligned

What you can test right now

  • Profile cognition with a short, non-diagnostic quiz.
  • Collect behavioral signals from micro tasks.
  • Render the same lesson in multiple structures.
  • Compare learner comfort, speed, and recall.

Proof of concept modules

These modules are live today. Together they show how NeuroStell profiles learners, maps cognitive needs, and adapts content without changing the underlying curriculum.

Profile

Cognitive constellation quiz

Probabilistic profiling from self-report preferences. Not a diagnosis.

Signals

Behavioral micro tasks

Short browser games capture attention shifts, reaction time, and memory bias.

Adapt

Perceptual rendering layer

Switch the structure of the same lesson: skimmer, stepper, deep, visual, or Socratic.

Multisensory

Focus soundscapes

Optional audio layers that reduce fatigue and support pacing.

Adaptive rendering preview

Same content, different structures. Use the quick switcher below to preview how NeuroStell can reduce cognitive friction without changing the information itself.

Why it matters

Institutions see drop-offs when learners cannot match their attention rhythm to the delivery format. Adaptive rendering targets this mismatch without demanding extra time from educators.

  • Skimmer: fast orientation and low load.
  • Stepper: guided flow for working memory support.
  • Deep: detailed explanation for mastery.

How the platform works

The PoC maps to the production workflow: profile learners, build interpretable neuro-constellations, adapt content, and deliver through any LMS.

Step 1 Profile

Questionnaires and behavioral tasks infer cognitive preferences.

Step 2 Map

Neuro-Constellations express attention, memory, and sensory load.

Step 3 Adapt

Rule-based logic plus LLMs render the optimal format and pacing.

Step 4 Learn

Deliver through web or LMS with SCORM/xAPI compatibility.

Validation and safeguards

NeuroStell is designed as an educational support tool. It does not diagnose. Validation focuses on learning gain, retention, and cognitive load proxies.

Research loop

Controlled pilots with institutions, followed by peer-reviewed studies.

Privacy by design

Minimal data, de-identified signals, optional exports, GDPR-ready.

Institution ready

White-label delivery, LMS integration, and accessible output formats.

Built by people who work across science, learning, and technology

NeuroStell is led by an interdisciplinary team with deep experience in learning science, product leadership, and AI infrastructure.

Sofia Gallego

Founder and CEO

Interdisciplinary scientist. Learning, complex systems, AI.

Jean‑Philippe M.C.

COO

20+ years product leadership. Serial entrepreneur in EdTech.

Ulysse Trin

CTO

Data scientist. Deep learning & AI/ML infrastructure.

Contact

Reach the team to discuss pilots, partnerships, or research collaboration.

Email:

sofiag.gallego@gmail.com

jpmdc@chastenay.co