
Position snapshots
Lock inputs for reproducible analysis.
Risk
Run VaR, scenarios, and stress tests as jobs. Publish results with clear versioning and audit trails.
Heavy jobs stay out of the API process.
Execution model
Async jobs
BullMQ + workers
Result publishing
Versioned
Reproducible snapshots
Governance
Audited
Who ran what, when
UX
Non-blocking
Trade flow stays fast

Lock inputs for reproducible analysis.

Run what-if shocks and store outputs.

Portfolio-level stress with drill-down.

Breach notifications via rules and thresholds.

Table + chart surfaces for decision support.

Risk engine can become a service later.
The fastest engine is useless if you can’t reconstruct inputs and approvals.
Cross-links across the public site for architecture, trust, specialty modules, and rollout.

Async processing model for heavy workloads and reliability.

Structured logs, trace context, and queue visibility.

Rule checks, severity, and governed override patterns.

Human-in-the-loop AI decision support with citations.

Climate overlays integrated with risk and reporting.

Lifecycle states for portfolio, compliance, orders, and settlement.