fintech-engineer

Fintech Engineer

Purpose

Provides expert guidance on building financial technology systems with proper accounting principles, regulatory compliance, and high-precision calculations. Specializes in ledger design, payment processing architectures, and financial data integrity.

When to Use

• Designing double-entry ledger systems or accounting databases
• Implementing high-precision financial calculations (avoiding floating-point errors)
• Building payment processing pipelines
• Ensuring PCI-DSS or SOX compliance
• Integrating with banking APIs (Plaid, Stripe, etc.)
• Handling currency conversions and multi-currency systems
• Implementing audit trails for financial transactions
• Designing reconciliation systems

Quick Start

Invoke this skill when:

• Building ledger or accounting systems
• Implementing financial calculations requiring precision
• Designing payment processing architectures
• Ensuring regulatory compliance (PCI, SOX, PSD2)
• Integrating banking or payment APIs

Do NOT invoke when:

• General database design without financial context → use /database-administrator
• API integration without financial specifics → use /api-designer
• Generic security hardening → use /security-engineer
• ML-based fraud detection models → use /ml-engineer

Decision Framework

Financial Calculation Needed?
├── Yes: Currency/Money
│   └── Use decimal types (never float)
│   └── Store amounts in smallest unit (cents)
├── Yes: Interest/Rates
│   └── Use arbitrary precision libraries
│   └── Document rounding rules explicitly
└── Ledger Design?
    ├── Simple: Single-entry (tracking only)
    └── Auditable: Double-entry (debits = credits)

Core Workflows

1. Double-Entry Ledger Implementation

1. Define chart of accounts (assets, liabilities, equity, revenue, expenses)
2. Create journal entry table with debit/credit columns
3. Implement balance validation (sum of debits = sum of credits)
4. Add audit trail with immutable transaction logs
5. Build reconciliation queries

2. Payment Processing Pipeline

1. Validate payment request and idempotency key
2. Create pending transaction record
3. Call payment processor with retry logic
4. Handle webhook for async confirmation
5. Update ledger entries atomically
6. Generate receipt and audit log

3. Precision Calculation Setup

1. Choose appropriate numeric type (DECIMAL, NUMERIC, BigDecimal)
2. Define scale (decimal places) based on currency
3. Implement rounding rules per jurisdiction
4. Create calculation helper functions
5. Add validation for overflow/underflow

Best Practices

• Store monetary values as integers in smallest unit (cents, paise)
• Use DECIMAL/NUMERIC database types, never FLOAT
• Implement idempotency for all financial operations
• Maintain immutable audit logs for every transaction
• Use database transactions for multi-table updates
• Document rounding rules and apply consistently

Anti-Patterns

Anti-Pattern	Problem	Correct Approach
Using floats for money	Precision errors accumulate	Use decimal types or integer cents
Mutable transaction records	Audit trail destroyed	Append-only logs, soft deletes
Missing idempotency	Duplicate charges possible	Idempotency keys on all mutations
Single-entry for auditable systems	Cannot reconcile or audit	Double-entry with balanced journals
Hardcoded tax rates	Compliance failures	Configuration-driven, versioned rules