Skip to content
[ PUBLIC TECHNICAL BRIEF · 2026 ]Working product

A capability system that changes only when the evidence changes.

The Operator Standard keeps the bar stable. The Adaptive Core changes the route only when accepted evidence changes what the learner has demonstrated. Every revision stays inspectable; attendance, tool usage, and self-identification move nothing.

Run the Operator Trial Follow the working product
MODEL-PORTABLE · PROVIDER-AWAREEVIDENCE-FIRSTRUBRIC-LEDENTERPRISE-BOUND
ADAPTIVE CORETRACE / 01
01 / SIGNAL

Your work, not a persona quiz

02 / SCORE

15 criteria → 6 dimensions

03 / COMPILE

Level + track + prerequisite-safe path

04 / PROVE

Artifact → rubric → graph → record

NEXT PATHVARIABLE DEPTH

fit(weak dimensions × curriculum track × level band)
→ prerequisite order → shipped capstone

ASSESS THE WORKCOMPILE THE PATHSHIP THE ARTIFACTVERIFY THE CLAIMREASSESS THE OPERATORADAPT WHAT COMES NEXT
[ FDE ROLE APPLICATION · INVESTOR VIEW ]

Build the training system once. Go deep by role.

LockedIn Labs is the implementation, consulting, and delivery firm. Its training platform is the practitioner-formation system; FDE is the first deeply specified role application inside it—not a new company, parallel product, or replacement brand.

The reuse thesis is visible in the product architecture today. Market adoption, field performance, and operating leverage still require measured engagements; none is inferred here.

FDE modules
08
Role domains
09
Evidence gates
10
DEPLOY stages
06
ONE SYSTEM · ONE DIRECTIONFORMATION → ROLE → METHOD → ENGAGEMENT
  1. 01 · SHARED CORE01

    LockedIn Labs training system

    Assessment, graph, labs, review provenance, and bounded credentials form the reusable operating substrate.

    FORMATION
  2. 02 · FIRST DEEP WEDGE02

    FDE role application

    8 modules and 9 domains specialize the shared system for forward-deployed engineering.

    ROLE
  3. 03 · FIELD EXECUTION03

    The DEPLOY Method

    6 evidence-ending stages turn the role standard into a repeatable field operating method.

    METHOD
  4. 04 · COMMERCIAL APPLICATION04

    Enterprise engagement

    A real workflow, named owners, client constraints, supervised evidence, and ownership transfer activate on contract.

    ENGAGEMENT
THE REUSABLE FORMATION LAYER

Every role application inherits the same evidence mechanics.

The role package changes what competent practice looks like. It does not fork identity, assessment infrastructure, graph resolution, lab execution, or proof provenance.

  1. ASSESS01

    Capability assessment

    One published judgment contract establishes the shared baseline.

  2. GRAPH02

    Prerequisite graph

    Role paths reuse dependency, routing, and evidence-gate mechanics.

  3. LABS03

    Executable practice

    The lab runtime, replay controls, and grading contracts are shared.

  4. REVIEW04

    Review provenance

    Every verdict preserves its reviewer, method, evidence, and limit.

  5. BOUNDARY05

    Qualification boundary

    Credentials derive from the graph; stronger claims require stronger proof.

WHAT IS FDE-SPECIFICAUTHORED IP
  • Role-specific capability standard
  • Advanced L2–L4 curriculum path
  • DEPLOY operating method
  • Field evidence and portability gates
  • FDE role targets and mission language
WHAT IS ENGAGEMENT-SPECIFICCONTRACT-ACTIVATED
  • The client workflow and outcome owner
  • Enterprise systems, controls, and constraints
  • A separately ratified client review standard
  • Named supervision and field observations
  • Client-owned handoff and operating artifacts
INVESTOR READING

FDE demonstrates that one shared training system can support a deeply authored role package without becoming a separate product stack. The next validation is commercial and operational: run the engagement, measure the evidence, and test whether another role reuses the same substrate.

Inspect the FDE standard Inspect the training system
01
[ The product thesis ]

The loop—not the library—is the product.

A content catalog can tell everyone what to watch. An adaptive capability system must observe what a person can do, prescribe the right work, judge the result, and use the new evidence to change the next decision.

  1. 01 / 07Implemented

    Observe

    Role, operating context, tools, a real prompt, an agent brief, and a recurring workflow.

  2. 02 / 07Implemented

    Diagnose

    Fifteen authored criteria roll into six capability dimensions, each scored 0–4 with evidence.

  3. 03 / 07Implemented

    Route

    Level, role-track fit, weak dimensions, prerequisites, and a proof-bearing capstone compile the path.

  4. 04 / 07Implemented

    Practice

    The learner works inside authored modules, labs, and a reusable playbook; cohort sessions activate only after roster and delivery coverage are confirmed.

  5. 05 / 07Implemented

    Verify

    Artifacts meet the module’s own rubric before they close a node or support a credential gate.

  6. 06 / 07Implemented

    Compare

    A reassessment shows dimension movement, level movement, and whether work was actually reperformed.

  7. 07 / 07Implemented

    Adapt

    A deterministic planning engine re-weights the next steps from goals, progress, reviewed artifacts, and reassessment evidence.

  8. Recursive learning means each proof changes what the system asks of you next.

02
[ Architecture ]

Authored knowledge. Runtime evidence. One governed judgment plane.

The architecture keeps curriculum IP separate from learner state, and keeps model judgment behind server boundaries. That makes the training platform fast to read, safe to personalize, and explicit about what a model did—or did not—decide.

CAPABILITY FOUNDRY · SYSTEM LAYER

Keep the core stable. Make frontier change traceable.

LockedIn Labs is the implementation, consulting, and delivery firm. It operates a training platform whose shared capability core is the AI Native Operator Framework; the Operator Standard fixes the bar. Capability Foundry turns sourced change into bounded packages; Fieldcraft practices them; Adaptive Core routes the individual; Evidence Chain preserves what the work can prove. The first authored lineage is inspectable; external practitioner ratification and continuous Foundry operation are not active.

CF-SYSTEM-01 · Capability FoundryINSIDE AI Native Operator Framework
LOCKEDIN LABS TRAINING · ONE PLATFORM
  1. 1CORE STANDARD

    AI Native Operator Framework

    Defines the stable capability standard: six dimensions, five levels, and the bar evidence must clear.

  2. 2CURATION

    Capability Foundry

    Defines the governed protocol for curating frontier change into versioned Capability Packages without creating a new product surface.

  3. 3TOPOLOGY

    Capability Graph

    Maps each package into prerequisite-aware curriculum nodes and the authored modules behind them.

  4. 4PRACTICE

    AI Fieldcraft

    Supplies the repeatable method for testing an unfamiliar capability under real constraints.

  5. 5ROUTING

    Adaptive Core

    Routes the individual through prerequisite-aware practice in response to recorded evidence.

  6. 6PROOF

    Evidence Chain

    Preserves the provenance, attempts, reviews, and verification behind every capability claim.

AUTHORED LINEAGE · CP-001 · V1.0.0

Networked shell agent release authorization

SOURCE → FIELDCRAFT → PACKET → GRAPH → RUBRIC

5 references → FCM-001 → L4 practice · L3 curriculum anchor · BME-06 → n-l3-3 → 4 dimensions / 100%

Editorial status authoredSource freshness CurrentPractitioner ratification Not activeSource renewal Scheduled
SYSTEM MAP / REQUEST TO RECORDSERVER BOUNDARIES PRESERVED
01EXPERIENCE

One product, four views

Public assessment method · learner instrument · expert review queue · employer-facing verifier

02APPLICATION

Server-first web platform

Next.js 16 · React 19 · TypeScript · route handlers · server components · Tailwind CSS v4

03JUDGMENT

Provider-portable, provider-aware model seam

Anthropic → OpenAI → Google → labeled deterministic fallback · provider provenance · schema validation

04EVIDENCE

Two data seams, kept separate

Versioned curriculum Repo · async learner Store · Postgres adapter for persistent deployments · offline memory adapter

05TRUST

Proof survives the interface

Opaque server sessions · scrypt passwords · authored rubrics · provenance stamps · public credential records

AUTHORED REPO · SYNC

Curriculum is versioned IP.

Levels, modules, rubrics, graph topology, and credential specs are authored in code and prerendered. A model cannot silently rewrite the standard it is scoring against.

LEARNER STORE · ASYNC

Evidence belongs to the operator.

Sessions, progress, assessments, artifacts, reviews, playbooks, enrollments, and credentials use real CRUD backed by Postgres when configured.

MODEL SEAM · VALIDATED

Models advise. Code decides.

Providers return structured judgments; application code validates the shape, clamps scores, applies gates, records provenance, and falls through safely on failure.

03
[ Inside the AI ]

The model never gets to invent the standard.

Versioned, published criteria define the standard. Configured models evaluate bounded evidence against those criteria. Deterministic application logic turns a validated verdict into placement, routing, progress, and proof.

INPUT / UNTRUSTED01

Read how the person works

  • Five bounded profile fields
  • A real prompt for a real task
  • An executable agent brief
  • A recurring workflow decomposition
JUDGMENT / STRUCTURED02

Score only the published rubric

  • 15 fixed assessment criteria
  • 0–4 score plus evidence note
  • Strict structured-output contract
  • Provider + model provenance stamp
DECISION / DETERMINISTIC03

Compile an explainable route

  • 6 dimension roll-up
  • Median-based L0–L4 placement
  • Inspectible role-track matching
  • Prerequisite-safe path to a capstone
FAIL-SAFE CONTRACT

Honest degradation beats synthetic certainty.

ASSESSMENT

If no configured frontier model returns a valid result, the heuristic is labeled and placement is capped at L1. It confirms participation—not advanced capability.

ARTIFACT REVIEW

The offline check can suggest a next step, but it cannot pass an artifact, close a graph node, or mint a credential. Proof requires a real judgment path.

04
[ Personalization ]

Every recommendation carries a reason. Every change needs evidence.

Personalization is not a demographic label or a chat interface. It is a set of bounded signals that change placement, sequence, feedback, and eventually the path itself.

Demonstrated exercises

L0–L4 placement and a six-dimension capability profile

Implemented

Role + operating context

The matched academy: foundations, engineering, agents, governed SDLC, leadership, or healthcare

Implemented

Two weakest dimensions

Which real curriculum modules rank highest inside the learner’s level band

Implemented

Graph prerequisites

The order of work; no recommendation appears before its dependencies

Implemented

Artifact rubric verdict

Whether a node closes and the specific correction the learner should make next

Implemented

Reassessment delta

The profile-over-time view and deterministic path recompilation, persisted as an inspectable revision ledger in the workspace

Implemented
RECURSIVE ENGINE / IMPLEMENTED

Reassess at meaningful gates. Blend the new profile with artifact evidence. Recompile only when the evidence warrants it.

The engine now recomputes an inspectable plan from the learner’s goal, latest assessment, graph progress, and bounded AI or expert review evidence—and the workspace persists that goal and appends each revision with the trigger and delta that produced it, without moving the authored credential standard or erasing history. An explicit reassessment workflow and a consented pilot remain ahead.

WATCH IT DECIDE8 PRECOMPILED ENGINE RUNS
ADAPTIVE-PATH/V1 · SYNTHETIC LEARNER · REAL ENGINE OUTPUT

Ships daily; has never operated an agent in production.

EVIDENCE TIMELINE
REVISION 01 · Initial build
+15 added✓ 5 newly complete

AI Native Developers path revision 1: 5/15 nodes complete; The AI-capable developer day is next. The route reflects the latest baseline, progress, and reviewed work.

✓ 4 earlier steps complete
Your personal AI operating systemL1 · 38 MIN
The AI-capable developer dayELECTIVEL1 · 31 MIN
The AI native dev environmentL2 · 37 MIN
Spec-driven development with agentsL2 · 47 MIN
Anatomy of an agentL2 · 31 MIN
Tools, MCP, and protocolsL2 · 36 MIN
Test-driven agentic codingL3 · 41 MIN
+ 4 further authored steps on this route
WHY THE NEXT STEP IS HERE

Elective depth on your AI Native Developers route — routed for this goal, outside every credential gate.

Both learners are synthetic and scripted; every plan, revision number, trigger, and delta above is the production engine's verbatim output over that script, compiled from the authored curriculum at build time. The engine never marks progress or grants mastery — those come only from reviewed work.

05
[ The end state ]

Expert is a standard of work—not a course-completion badge.

The objective is employable operating capability: people who can frame problems, build with agents, verify behavior, govern consequences, and explain the evidence to an enterprise team.

L0
AI Curious

See the machine clearly.

Strip away the magic. Build accurate mental models of what modern AI systems are, what they are not, and where they genuinely bite.

GATE · Planned: an unscripted foundations review, active after vetted examiner coverage.

L1
AI Capable

Run your day on AI.

Turn AI from a novelty into your default way of working. Design personal workflows, write prompts that behave like specifications, and measure what you save.

GATE · Ship two verified artifacts: a working personal workflow and a hardened prompt system.

L2
AI Practitioner

Apply it where you work.

Take the craft into your role — engineering, operations, clinical work, or the executive suite. Role-track modules with domain constraints and real stakes.

GATE · Ship a role-track capstone artifact. Planned: named practitioner review, active after vetted reviewer coverage.

L3
AI Native Operator

Build and run agentic systems.

The operator level. Compose multi-step agents, orchestrate fleets, evaluate behavior rigorously, and take systems from demo to dependable.

GATE · Ship a verified agentic system: working, evaluated, and documented.

L4
AI Native Leader

Transform the organization.

From operator to force multiplier. Strategy, governance, operating models, and the economics of AI transformation — for people who move companies, not just workflows.

GATE · Planned: defend a transformation plan in a live board-style review, active after vetted examiner coverage.

6Capability dimensions
15Assessment criteria
7Authored academies
42Proof-bearing modules
42Skill-graph nodes
7Credential specifications
BUILD

Production-shaped artifacts

Specs, workflows, eval suites, agent systems, control matrices, authorization packages, and transformation plans—not quiz scores alone.

DEFEND

Judgment under questioning

When vetted reviewer coverage is active, live reviews and practitioner spot-checks test whether the learner can explain tradeoffs, failure modes, controls, and rollback stories.

PUBLISH

Evidence an employer can inspect

Credential gates derive from the graph and resolve to public records that expose requirements and linked work instead of a decorative badge.

06
[ Open ecosystem ]

Portable at the core. Provider-aware at the edge.

The durable standard travels across changing models and platforms, but implementation never ignores provider-specific behavior or constraints. Official-source supplements can add platform fluency as a clearly labeled layer updated on the provider’s cadence.

LANE 01Implemented
THE DURABLE STANDARD

LockedIn core

Problem framing, decomposition, agent design, evals, governance, evidence, and enterprise delivery. These skills survive a model or platform change.

LANE 02Roadmap
OFFICIAL-SOURCE SUPPLEMENTS

Platform labs

Claude and Anthropic workflows, OpenAI APIs and evals, Amazon and Google deployment patterns, and other platform-specific operating labs.

LANE 03Activation gate
PARTNERSHIP OR PUBLIC SPEC

Credential mapping

Map official exam objectives to proof-bearing practice where terms allow. LockedIn credentials never imply vendor endorsement or substitute for an official certification.

CURATION RULE

Supplemental never means copied.

SOURCE

Link to official, permissioned material and record its version.

PRACTICE

Add an original lab that turns documentation into a working artifact.

VERIFY

Assess the transferable skill and the platform-specific implementation separately.

07
[ Who this is for ]

One evidence system. Four reasons to care.

The mechanism stays consistent; the value changes by audience. Learners need direction, practitioners need a credible standard, enterprises need capability evidence, and partners need a responsible way to extend the system.

LEARNER

A path that starts with your work

Know where you stand, why each module was chosen, what proof closes the gap, and how your capability is moving over time.

Create account to assess
PRACTITIONER

A standard worth teaching against

Author real methods, review consequential work, mentor emerging operators, and help define what competent practice looks like.

See the practitioner model
ENTERPRISE

Capability you can inspect

Scope a team capability program around operating work and define the evidence leaders should inspect instead of relying on attendance.

Explore enterprise delivery
PARTNER

An ecosystem, not a walled garden

Bring official source material, specialist labs, platforms, and hiring pathways into a shared proof-of-work standard.

Start a partner conversation
[ THE NEXT CONVERSATION ]

Don’t review the promise. Walk through the system.

We’ll show the assessment, the scoring contract, a personalized path, the learner workspace, artifact review, and the public proof record—then decide what to build together.

45-MINUTE WORKING SESSION
  1. 01 · PRODUCT WALKTHROUGH
  2. 02 · TECHNICAL ARCHITECTURE
  3. 03 · ADAPTIVE ROADMAP
  4. 04 · PILOT OR PARTNERSHIP FIT
Request a walkthrough Inspect product proof