TradAI Architecture¶
Welcome to the architecture documentation. Start with the overview, then dive into specifics.
Start Here
Read Architecture at a Glance for a complete system diagram and reading paths.
Architecture Documents¶
Infrastructure¶
| Document | Description |
|---|---|
| System Design | Component inventory, communication matrix |
| VPC & Networking | Subnets, security groups, NAT, endpoints |
| Security | Cognito, IAM, WAF, encryption, audit |
| Services | ECS services, Lambda functions, Docker Compose |
| Step Functions | Backtest and retraining workflows |
| Cost Analysis | Cost breakdown and optimizations |
| Pulumi Code | 4-stack IaC architecture |
| Canonical Config | Configuration source of truth |
| Live Trading | Live trading architecture (aspirational) |
| ML Lifecycle | Training, drift detection, retraining |
Note: Document 08 (Implementation Roadmap) was archived after project completion.
Application Architecture¶
| Document | Description |
|---|---|
| Data Flows | End-to-end data movement |
| State Machines | Job, trading, promotion lifecycles |
| Error Handling | Retry, circuit breaker, DLQ |
| Observability | Logging, metrics, health, tracing |
| Configuration | Infra vs app config systems |
| CI/CD Pipeline | GitHub Actions workflows |
| Testing Strategy | Test suites, markers, coverage |
| Deployment Pipeline | Bootstrap, Lambda, ECS deployment |
Technical Design¶
| Document | Description |
|---|---|
| tradai-common | Base classes, AWS utilities |
| tradai-data | Data fetching, ArcticDB |
| tradai-strategy | Strategy framework |
| Backend | API gateway design |
| Data Collection | Market data service |
| Strategy Service | Backtest execution |
| MLflow | Experiment tracking |