Expanding the Platform

Relevant source files

The following files were used as context for generating this wiki page:

• README.md
• backend/src/app.ts
• infrastructure/aws/cdk/lib/atlasclean-stack.ts
• infrastructure/terraform/main.tf
• backend/src/infrastructure/database/prisma/schema.prisma
• backend/src/infrastructure/services/AtlasAIService.ts
• packages/atlas_design_system/pubspec.yaml

AtlasCleanAI Enterprise provides a modular foundation for scaling an AI-powered cleaning services ecosystem. The platform supports multiple client applications, scalable cloud infrastructure, and an intelligence layer for operational optimization.

Expanding the platform involves adding new client applications, extending backend services via use cases, and scaling the infrastructure using Infrastructure-as-Code (IaC) templates. The monorepo structure ensures that shared logic, design systems, and API clients remain consistent across all new modules.

Sources: README.md:1-25, backend/src/app.ts:1-50

Multi-Platform Extension

The platform supports expansion through a monorepo structure that facilitates the addition of new applications. Developers can create new Flutter applications in the apps/ directory while consuming shared logic from the packages/ directory.

Shared Design and API Packages

Consistency across new platforms is maintained via:

• atlas_design_system: A Flutter package containing brand colors, typography tokens, and pre-styled UI components.
• atlas_api_client: A Dart package for standardized HTTP communication, JSON parsing, and authenticated requests.

Sources: README.md:126-146, packages/atlas_design_system/pubspec.yaml:1-15

Client Ecosystem

The architecture is designed to support four primary interfaces:

1. Customer App: For service booking and lifecycle management.
2. Cleaner App: For onboarding, scheduling, and job execution.
3. Management App: For workforce and billing oversight.
4. Web Portal: For administrative tasks and reporting.

Sources: README.md:173-180

Backend Scalability

The backend expansion relies on a clean architecture where new features are implemented as Use Cases and exposed through Controllers.

The diagram shows the request flow from various client applications through the load balancer to the application logic and persistence layers.
Sources: backend/src/app.ts:165-190, infrastructure/aws/cdk/lib/atlasclean-stack.ts:220-240

Domain Model Extension

New service types or business entities are added by updating the Prisma schema and generating the client.

Entity	Description	Purpose
User	Stores identity and roles	Supports CUSTOMER, ADMIN, CLEANER, and MANAGER roles.
Cleaner	Stores professional data	Manages skills, availability, and background checks.
Service	Defines offerings	Configures base price, duration, and categories.
Booking	Manages transactions	Links users, cleaners, and services with status tracking.

Sources: backend/src/infrastructure/database/prisma/schema.prisma:35-130

AI-Powered Intelligence Layer

Platform expansion includes enhancing the AtlasAIService to provide deeper operational insights. This service acts as a bridge between the core application and AI models.

AI Service Capabilities

The intelligence layer currently performs the following:

• Schedule Optimization: Ranks cleaners based on skills, rating, and location.
• Demand Prediction: Analyzes postal codes and service types to predict capacity needs.
• Dynamic Pricing: Calculates price adjustments based on demand multipliers and rush hour factors.
• Cleaner Recommendations: Identifies the best-fit professional for specific job requirements.

Sources: backend/src/infrastructure/services/AtlasAIService.ts:35-125

This sequence illustrates the integration of AI-driven decision-making within the booking workflow.
Sources: backend/src/infrastructure/services/AtlasAIService.ts:40-65

Infrastructure and Deployment

Expansion into new regions or environments is managed through AWS CDK or Terraform. Both implementations provide isolated network tiers and managed services.

Infrastructure Components

Component	Implementation	Scaling Strategy
Compute	AWS ECS Fargate	Auto-scales based on CPU utilization (target 60%).
Database	Amazon RDS PostgreSQL	Multi-AZ deployment in production with storage auto-scaling.
Caching	Amazon ElastiCache Redis	Primary/Replica clusters with automatic failover.
Networking	VPC with 3 Tiers	Public, Application, and Data subnets across 2 AZs.

Sources: infrastructure/aws/cdk/lib/atlasclean-stack.ts:35-150, infrastructure/terraform/main.tf:40-100

Multi-Environment CI/CD

The platform utilizes GitHub Actions for staged deployments:

1. CI Pipeline: Validates code quality and security.
2. Staging: Deploys to a non-production AWS account for validation.
3. Production: Gated deployment using environment protections and AWS OIDC.

Sources: README.md:94-105, infrastructure/aws/cdk/package.json:7-10

Conclusion

Expanding the AtlasCleanAI platform requires a coordinated effort across the shared Flutter packages, the Node.js backend use cases, and the infrastructure-as-code definitions. The architecture supports horizontal scaling of the API tier and vertical scaling of the data tier, while the AI Intelligence Layer provides the flexibility to add complex optimization logic without disrupting the core domain.