isA Data
Data-as-a-Service platform combining intelligent RAG with enterprise data lake management.
Overview
isA Data provides a unified platform for knowledge management and structured data analytics:
| Capability | Description |
|---|---|
| Intelligent RAG | 7 RAG patterns for knowledge retrieval |
| Data Lake | Zone-based storage (Raw → Curated → Gold) |
| Data Fabric | Natural language query interface |
| Multi-Modal | PDF, Image, Audio, Video processing |
| Data Products | Self-service analytics with quality guarantees |
Latest Updates (2026-05-17)
- Framework data-product marketplace now supports product search, product detail, and subscribe flows.
- Certification report APIs expose product readiness evidence for marketplace listings.
- Policy evaluation gates data-product status transitions and certification decisions.
- Support and trust cards surface operational and governance context in marketplace results.
- Canonical product IDs and framework schema bundle/versioning are now part of the data-product contract.
Architecture
┌─────────────────────────────────────────────────────────────────────────────┐
│ API LAYER │
│ ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐ │
│ │/digital │ │ /fabric │ │/ingest │ │ /etl │ │/catalog │ │
│ └─────────┘ └─────────┘ └─────────┘ └─────────┘ └─────────┘ │
└─────────────────────────────────────────────────────────────────────────────┘
│
┌─────────────────────────────────────────────────────────────────────────────┐
│ SERVICE LAYER │
│ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ │
│ │ Digital │ │ Data Fabric │ │ Data Infra │ │Data Product │ │
│ │ Service │ │ Service │ │ Service │ │ Service │ │
│ │ │ │ │ │ │ │ │ │
│ │ • 7 RAG │ │ • NL Query │ │ • Ingestion │ │ • Register │ │
│ │ • Store │ │ • SQL Gen │ │ • ETL │ │ • Quality │ │
│ │ • Search │ │ • Schema │ │ • Zones │ │ • Contracts │ │
│ │ • Evaluate │ │ • Metadata │ │ • CDC │ │ • SLA │ │
│ └─────────────┘ └─────────────┘ └─────────────┘ └─────────────┘ │
└─────────────────────────────────────────────────────────────────────────────┘
│
┌─────────────────────────────────────────────────────────────────────────────┐
│ INFRASTRUCTURE LAYER │
│ ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐ │
│ │ Qdrant │ │ MinIO │ │PostgreSQL│ │ Neo4j │ │ NATS │ │
│ │ (Vector)│ │ (S3/Lake)│ │ (Meta) │ │ (Graph) │ │ (Events)│ │
│ └─────────┘ └─────────┘ └─────────┘ └─────────┘ └─────────┘ │
└─────────────────────────────────────────────────────────────────────────────┘Core Services
Digital Service (RAG/Knowledge)
Intelligent retrieval-augmented generation with 7 RAG patterns:
| Pattern | Description | Best For |
|---|---|---|
| Simple RAG | Basic vector search + LLM | General Q&A |
| CRAG | Corrective RAG with validation | Fact-checking |
| HyDE | Hypothetical Document Embeddings | Sparse data |
| Graph RAG | Knowledge graph enhanced | Entity relationships |
| RAG Fusion | Multi-query fusion | Complex questions |
| Self-RAG | Self-reflective generation | High accuracy |
| RAPTOR | Recursive summarization trees | Long documents |
Data Infrastructure Service
Delta Lake management with zone-based architecture:
Raw Zone Curated Zone Gold Zone
┌─────────┐ ┌─────────┐ ┌─────────┐
│Immutable│ ──▶ │ Cleaned │ ──▶ │Aggregated│
│ Source │ │Validated│ │Optimized │
│ 365 days│ │ 180 days│ │ 90 days │
└─────────┘ └─────────┘ └─────────┘Data Fabric Service
Natural language to SQL translation for self-service analytics:
- Schema-aware query generation
- Metadata embedding for discovery
- Result formatting and visualization
Data Product Marketplace
Framework data products can be discovered, certified, and subscribed to through the catalog API.
| Method | Route | Purpose |
|---|---|---|
GET | /api/v1/data/catalog/framework/marketplace/products | Search marketplace products |
GET | /api/v1/data/catalog/framework/marketplace/products/{product_id} | Read product detail |
POST | /api/v1/data/catalog/framework/marketplace/products/{product_id}/subscribe | Subscribe to a product |
GET | /api/v1/data/catalog/framework/products/{product_id}/certification | Read certification state |
POST | /api/v1/data/catalog/framework/products/{product_id}/certification/report | Submit certification evidence |
GET | /api/v1/data/catalog/framework/products/{product_id}/certification/report | Read certification report |
POST | /api/v1/data/catalog/framework/products/{product_id}/status | Request or apply a product status transition |
File Processing
Multi-modal content extraction:
| Type | Capabilities |
|---|---|
| Text, tables, images, OCR | |
| Image | AI analysis, descriptions, metadata |
| Audio | Transcription (Whisper) |
| Video | Frames, scenes, transcript |
Quality Metrics
| Metric | Target |
|---|---|
| RAG Faithfulness | > 0.8 |
| RAG Relevancy | > 0.75 |
| NL Query Latency | < 2s |
| Data Freshness (Raw) | < 1h |
| Data Freshness (Gold) | < 24h |
Infrastructure Dependencies
| Service | Purpose | Port |
|---|---|---|
| Qdrant | Vector database | 6334 |
| MinIO | Object storage (S3) | 9000 |
| PostgreSQL | Metadata store | 5432 |
| Neo4j | Graph database | 7687 |
| NATS | Event streaming | 4222 |
| isA Model | Embeddings & LLM | 8082 |
Next Steps
- Quick Start - Get started
- RAG Patterns - 7 RAG modes
- Data Lake - Zone management
- Data Fabric - NL queries
- File Processing - Multi-modal