Case Study / Noosaga

Designing an atlas for difficult subjects.

Noosaga is a knowledge atlas built to orient people before the details take over. It brings together information architecture, product design, frontend build, and visible trust cues.

Role Founder, design, frontend Focus Orientation before detail Surfaces Timelines, comparisons, maps Constraint Trust without flattening nuance

Problem

Most learning tools lose the big picture just when the user needs it.

Search gives you one result. Textbooks give you one route. AI often gives you one fluent answer. None of them reliably show the shape of the field first.

Search

Fast local answers, weak global structure.

Textbooks

One authored route, limited comparison space.

Glossaries

Terms without dependency or argumentative shape.

AI output

Fluent summaries that can hide provenance and uncertainty.

Portrait of Axel Pond

Role

Axel Pond

Founder. Product direction, information architecture, interface design, and frontend build.

Scope

Framing, structure, routes, view logic, visual language, and implementation.

Constraint

The interface has to stay calmer than the idea. If it gets too loud, orientation breaks.

System

Four design problems had to be solved together.

Noosaga only works if structure, views, trust cues, and interface behavior reinforce each other.

02

Views

Different questions need different views: overview, comparison, concept tracing, and guidance.

03

Trust

If AI participates in the pipeline, provenance has to remain visible and legible inside the product.

04

Frontend craft

Motion, hierarchy, responsiveness, and state transitions have to support orientation rather than compete with it.

Interaction thesis

Each view should answer a different question.

The overview helps you get bearings. Comparison helps you judge competing schools. Concept maps show dependencies. Provenance shows how to read the material.

Overview Orient the field Comparison Evaluate rival schools Concept map Trace dependencies Provenance Make trust legible

Field spine first

Early screens privilege bearings and route choices over feature density.

Comparison as a first-class view

Schools and frameworks should be inspectable side by side instead of buried in sequential prose.

Concepts as a map

Users should be able to trace relationships instead of collecting disconnected glossary entries.

Trust inside the interface

Source state, inference, and uncertainty need to be visible in the reading flow, not hidden in notes.

Decisions

The main product decisions.

  • Keep the top-level frames stable so users always know where they are.
  • Separate views by reasoning mode rather than by content type alone.
  • Favor legible hierarchy over maximal density.
  • Treat provenance as interface content, not back-office metadata.

Engineering

What that means in code.

  • The content model has to power multiple views without losing semantic structure.
  • State needs to survive route changes, filtering, and responsive layout shifts.
  • Motion should clarify transitions rather than act as decoration.
  • The public product and portfolio site can share visual DNA while serving different constraints.

Proof Points

What this project is proving.

Design

A product can teach orientation.

The interface is treated as a thinking tool, not just a content wrapper.

Trust

AI-assisted structure can stay legible.

Trust depends on showing sources, inference, and uncertainty clearly.

Build

The visual layer can stay expressive.

The frontend is judged by how well it preserves clarity under real content and state.

Next

Want the live product, the homepage proof, or the writing around it?

Noosaga is the live artifact. The homepage is the short argument. The writing gives the broader worldview.