MIKIAS ABERA — SOFTWARE DEVELOPER & PRODUCT BUILDERIDEAS → EXPERIENCES

Curiosity.
Craft.Code.

I turn interesting questions into
beautiful, useful things.

I build quality software and experiences.
AI expands what I can make. Craft shapes the result.
Selected work
A sunlit field beneath an immense tree, from The golden field
NOETICMAP / SOFTWARE FOR EXPLORING CONSCIOUSNESS
SOFTWAREDESIGNAI SYSTEMSCURIOSITY
01 / SELECTED PRODUCTS

From the hard problem
to the finished product.

01 — NOETICMAPRESEARCH PLATFORM · AI EXTRACTION · EXPERIENCE DESIGN

An atlas of
consciousness.

I built NoeticMap to make experiences, research, and wisdom traditions easier to explore. Behind the interface is a substantial processing system that turns scattered documents into structured knowledge.

Explore NoeticMap
A luminous city imagined for the NoeticMap experience
NoeticMapA world of questions. A place to explore.

Research, structured.

Discovery, triage, and extraction turn papers into findings, methods, cases, and insights.

Knowledge, connected.

Semantic search and structured records help people explore experiences and research.

A complete experience.

Interface design, visualizations, audio, and film give people different ways into the subject.

65,074Papers discovered
2,233Papers fully extracted
10,063Structured key insights

A documented extraction run.
A snapshot of the work, not live totals.

WITHIN NOETICMAP

Another way into the questions.

Three AI-assisted films, created as part of the wider product experience.

02 — PERMITCHECKDOCUMENT AI · RULES ENGINE · PRODUCT DESIGN
A precise architectural model above drafting plans, made from pale concrete and frosted blue glassDRAWINGS → UNDERSTANDING
PermitCheckDOCUMENT INTELLIGENCE
HOW IT WORKS
1
Read the packageDrawings, forms & dimensions
2
Resolve the propertyAddress-specific zoning rules
3
Compare & explainReadiness report with cited rules
Documents become decisions.
FROM DENSE DOCUMENTS TO A CLEAR NEXT STEP

Less guesswork.
More readiness.

PermitCheck reads Toronto residential permit packages, extracts the proposed dimensions, and compares them with the zoning rules that apply to the actual property.

  • Multi-pass vision extraction from drawings and documents.
  • Address-specific zoning resolution and computed comparisons.
  • Cited findings, with manual review where a value can’t be confirmed.
Explore PermitCheck
03 — ENACTEDDATA PIPELINES · CHANGE DETECTION · AI SUMMARIZATION
A CONTINUOUS PIPELINE FOR CANADIAN LAW

Know what changed.
Understand why it matters.

Enacted turns changes to Ontario and federal law into something people can follow. The pipeline watches official sources, computes the differences, and produces readable explanations.

  • Official e-Laws data and Justice Canada’s XML repository.
  • Exact version comparisons before AI enters the process.
  • Grounded summaries linked back to the underlying changes.
Explore Enacted
Sculptural ivory document pages with one vermilion sheet, representing a change between versionsVERSIONS → CLARITY
EnactedCHANGE INTELLIGENCE
HOW IT WORKS
1
Track official sourcesOntario e-Laws & federal XML
2
Compute the changeExact differences between versions
3
Explain what changedAI-assisted plain-English summaries
Changes become understanding.
04 — BLINDFOLD LABAUDIO-FIRST SOFTWARE · PERCEPTION · INTERACTION DESIGN
A sculptural black blindfold suspended between translucent red, blue, and gold geometric objects
BLINDFOLD LABAN EXPERIMENT IN EXPERIENCE DESIGN
A DIFFERENT WAY TO INTERACT

Less looking.
More sensing.

Explore Blindfold Lab

What happens when
the screen isn’t the interface?

An audio-guided practice platform designed for use with your eyes closed. Voice guidance, keyboard controls, and swipe interactions carry the session. The design challenge is making every step understandable without looking.

VOICE GUIDANCECOLORS & SHAPESEYES-FREE CONTROLS
02 / BEHIND THE INTERFACE

Complex systems.
Useful experiences.

AI is part of how I build, from developing software to processing information at scale. The interesting work is making the whole system useful: the data, the models, the checks, and the experience at the other end.

INSIDE THE NOETICMAP EXTRACTION PIPELINESELECT A STAGE ↓

Turn unstructured text into useful data.

A scan pass and a deeper extraction pass turn documents into findings, cases, methods, and insights. I used local Qwen models on Apple Silicon for deep extraction, alongside API models for other stages.

IN

Selected documents and text chunks

OUT

Structured findings, cases, methods, and insights

Two-phase extraction · local inference · schema-shaped output
01

Choose the right work for AI.

Models help read, classify, extract, and explain. Calculations and exact comparisons belong in code.

02

Measure what matters.

Use evals and source checks to improve the product. Reliability is part of the craft.

03

Finish the experience.

The pipeline is only part of it. Make the result clear, responsive, and a pleasure to use.

THE PERSON BEHIND THE PRODUCTS

A developer.
An enduring curiosity.

I’m Mikias, a software developer passionate about building quality products and experiences. I care about the engineering beneath the surface and the way a product looks, feels, and works.

AI helps me build more ambitiously: from writing software to extracting knowledge from large collections of documents. Design, judgment, and attention to detail shape what I do with it.

THE QUESTIONS BEHIND THE WORK

Consciousness.
Experience. What it means to be human.

Consciousness is a deep personal interest. I’m drawn to questions about the nature of experience, the mind, and what we can learn from people’s extraordinary accounts. NoeticMap is one way I explore that curiosity through something I can build and share.

Explore those questions with me
Let’s talk products, ideas, or consciousness
GET IN TOUCH

Have something
in mind?

Tell me what you’re building, or send a question.

mikias@abera.io