boc/docs/design/LANDVEX_ENGINE_ARCHITECTURE_v1.0.md

LandveX Engine Architecture v1.0

Platform Overview

LandveX is not a collection of services. It is a platform of autonomous engines with clear responsibilities.

Reality Engine
  ↓
Knowledge Engine
  ↓
Decision Engine
  ↓
Mission Engine
  ↓
Economic Engine
  ↓
Learning Engine

Platform Orchestrator coordinates engines via events. Not AI. Not Dashboard. Just orchestration.

---

Engine 1 — Reality Engine

Responsibility: Transform reality into verified observations.

Input:

• quiXzoom (phone)
• Drones
• Cameras
• Video
• GIS
• Sensors

Output:

• Observation
• Artifact
• Metadata

Domain objects created:

• FieldSession
• Mission
• Observation
• ArchiveArtifact

Events produced:

• ObservationCreated
• MissionCompleted
• ArtifactRegistered

KPIs:

• Observations per hour
• Upload success rate
• Metadata completeness

---

Engine 2 — Knowledge Engine

Responsibility: Understand what observations mean.

Input:

• Observation
• ArchiveArtifact

Output:

• Evidence
• Finding
• KnowledgeArtifact
• Ontology entries
• Taxonomy classifications
• Object relations
• Knowledge Graph updates

Domain objects created:

• Evidence
• Finding
• KnowledgeArtifact

Events consumed:

• ObservationCreated

Events produced:

• EvidenceCreated
• FindingCreated
• KnowledgeExtracted

KPIs:

• Knowledge extraction accuracy
• Ontology coverage
• Taxonomy completeness

---

Engine 3 — Decision Engine

Responsibility: Make knowledge actionable.

Input:

• Finding
• Evidence
• Knowledge Graph

Output:

• Decision
• Priority
• Confidence
• Consequence
• Recommendation

Domain objects created:

• DecisionCase
• Decision
• Review

Events consumed:

• FindingCreated

Events produced:

• DecisionCreated
• DecisionApproved
• DecisionRejected

KPIs:

• Decision accuracy
• Time to decision
• Review completion rate

---

Engine 4 — Mission Engine

Responsibility: Determine what data to collect next.

Input:

• KnowledgeGap
• Hotspot
• Contradiction
• Customer requests
• Budget
• SLA

Output:

• Mission
• Reward
• Priority
• Coverage

Domain objects created:

• Mission
• KnowledgeGap
• Hotspot
• Contradiction

Events consumed:

• KnowledgeGapIdentified
• ContradictionDetected

Events produced:

• MissionCreated
• MissionAssigned
• MissionCompleted

KPIs:

• Knowledge gap closure rate
• Mission success rate
• Coverage improvement

---

Engine 5 — Economic Engine

Responsibility: Manage budgets, credits, and ROI.

Input:

• Budget
• Mission
• VerifiedDelivery

Output:

• Credit
• Settlement
• ROI
• RewardPricing
• CostAllocation

Domain objects created:

• Credit
• CreditBudget
• IntelligenceLedgerEntry

Events consumed:

• MissionCompleted
• DecisionApproved

Events produced:

• CreditsAllocated
• PaymentSettled
• RoiCalculated

KPIs:

• Cost per decision
• Knowledge return per credit
• Budget utilization

---

Engine 6 — Learning Engine

Responsibility: Continuously improve all engines.

Input:

• Outcome
• Decision
• Reality (ground truth)

Output:

• Better models
• Better rules
• Better missions
• Better pricing

Events consumed:

• DecisionApproved
• MissionCompleted
• OutcomeVerified

Events produced:

• ModelUpdated
• RulesRefined
• PricingAdjusted

KPIs:

• Model improvement rate
• Prediction accuracy gain
• Cost reduction

---

Platform Orchestrator

Responsibility: Coordinate engines via events.

Not AI. Not Dashboard. Just orchestration.

Example flow:

ObservationCreated
  ↓
Knowledge Engine (extract knowledge)
  ↓
Decision Engine (create decision)
  ↓
Mission Engine (check for gaps)
  ↓
Economic Engine (calculate ROI)

Rules:

• Engines never call each other directly
• All communication via events
• Orchestrator decides which engine runs next
• Each engine is independently versioned

---

Dashboard

Dashboard never shows engines.

Dashboard shows results:

AREA
  Score
  ↓
  Top Decisions
  ↓
  Hotspots
  ↓
  Business Impact
  ↓
  Recommended Actions

---

Intelligence Lab

Intelligence Lab is the development environment for engines.

Test:

• Knowledge Engine
• Decision Engine
• Learning Engine

Not:

• Whole platform
• Production data
• End-to-end flows

---

Engine Contracts

Each engine MUST define:

Contract	Description
Responsibility	What this engine does
Input	What it consumes
Output	What it produces
Domain Objects	What it can create/modify
Events In	Events it listens to
Events Out	Events it produces
KPIs	How quality is measured

---

Development Principles

1. Each engine independently versioned
2. Engines never call each other directly
3. All communication via events
4. Orchestrator decides flow
5. Dashboard shows results, not engines
6. Intelligence Lab tests engines, not platform

---

Related Documents

• ADR-011: Four-Layer Data Architecture
• ADR-012: Economic Engine
• LANDVEX_PLATFORM_ARCHITECTURE.md