About

About Mariana

I build the structure underneath trusted systems, companies, public thinking, and original work.

Mariana Lungu seated on a panel, hands folded, audience visible in the background

Origin

The work changed form. The standard stayed.

I care about what sits underneath the visible result: where it came from, what changed, who made the decision, and whether another person can understand it later.

That attention began in federal data work and now runs through Archetype Core, The Ferne, and everything I choose to make public.

Outside the work, I am a mother, a reader, and a painter. I am drawn to patterns, quiet materials, and the discipline required to make something lasting without making it loud.

01

Environment

I learned to build where almost right was not good enough.

For nearly six years, I worked on federal data systems tied to immigration cases, benefits processing, congressional inquiries, and records that affected real people. The visible output mattered, but the handoffs, lineage, and decisions underneath it mattered more.

02

Lesson

That work changed what I look for in every system.

I look for the source, the transformation, the judgment, and the record that remains after the screen changes. Trust is not a feeling added at the end. It is something the architecture has to preserve from the beginning.

03

Archetype Core

I founded a technical practice around that standard.

Archetype Core helps regulated teams build AI and data systems that can be inspected after launch: cited answers, reliable pipelines, visible lineage, governance, and decisions another person can reconstruct.

04

The Ferne

The same eye belongs in another form.

The Ferne is the art house. It holds the part of my work concerned with material, memory, restraint, and permanence. The medium changes, but the attention to structure and what remains does not.

Public thinking

I make the work visible while I am building it.

I write about provenance, trusted AI systems, data architecture, and the decisions underneath reliable output. Beyond the Prompt carries the arguments. YouTube shows how the systems are built.

Bring me in when the work has to stand up later.

Work with me