How TLDT works

TLDT uses one reviewed Career Memory across private preparation and optional public representation. The owner controls the publication boundary; every workflow reads an immutable version rather than a live draft.

The product sequence

Start from owner-supplied evidence

Add CVs, notes, transcripts, project records, and other material you own and can review.

Publish a reviewed Memory version

Choose which documents are Public, resolve warnings, publish an immutable version, and activate it.

Check a real role

Fast Check compares one vacancy with published evidence and distinguishes a real gap from missing stored proof.

Rehearse or test representation

Manual Simulate tests your answer. Auto Simulate tests what the pinned AI profile can answer from the same reviewed context.

Share only when useful

A public profile or saved snapshot exposes only the approved representation, never the owner workspace.

One Memory, several workflows

Workflow	Immediate job	Memory authority
Fast Check	Map evidence to one real vacancy and identify questions to validate.	Active published version at session creation.
Manual Simulate	Rehearse how you would answer a role-specific interview.	The session can exist without published Memory; you write every candidate answer.
Auto Simulate	Test how the AI profile represents you under interview pressure.	Published version pinned when the session is created.
Owner Preview	Test public-style answers privately before sharing.	Current active published version.
Public Chat and CV	Let a visitor explore approved professional context.	Current active published version.
Saved shares	Give a visitor a stable view of one completed resource.	Visitor-safe snapshot created by the owner.

The Memory boundary

Owner-supplied Raw evidence can support reviewed Derived Memory, which becomes available to workflows only after explicit publication — Evidence becomes public only through explicit review and publication

State	What it contains	Can public or Auto workflows use it?
Raw documents	Owner-supplied evidence stored with server-owned Properties.	Only selected Public excerpts included in the active published version.
Derived documents	Concise operational Memory linked to Raw sources.	Only reviewed Public documents included in the active version.
Draft workspace	Editable Raw and Derived catalogs.	No. Draft edits never change runtime Memory.
Published version	Immutable, validated projection selected by the owner.	Yes, when active or pinned to a session.

What improves automatically — and what does not

TLDT can analyze Properties and generate a Private Derived candidate from selected Raw sources. Those outputs remain drafts until you review and save them. A completed simulation or Chat can be saved to Raw Memory through an explicit owner action. TLDT does not automatically extract a Memory Delta, rewrite Derived Memory, reconcile contradictions, or publish new facts.

Representation limits

The AI can explain supported professional context and, when the owner allows it, label interpretation or general knowledge. It cannot invent personal history, expose Private Memory, apply for a role, negotiate, declare interest or availability, or make commitments for the owner.

Use TLDT to make missing proof visible. A gap in published Memory is a reason to review evidence or ask the owner, not permission to guess.

Start here

Prepare for a role

Career Memory

Represent and share

Settings and privacy

How TLDT works

The product sequence

One Memory, several workflows

The Memory boundary

What improves automatically — and what does not

Representation limits

​The product sequence

​One Memory, several workflows

​The Memory boundary

​What improves automatically — and what does not

​Representation limits

The product sequence

One Memory, several workflows

The Memory boundary

What improves automatically — and what does not

Representation limits