Validate virtual credit card transactions from encoded IDs
Company: Capital One
Role: Machine Learning Engineer
Category: System Design
Difficulty: medium
Interview Round: Onsite
You are designing logic for a **virtual credit card** product.
## Part 1: Product reasoning
Explain key benefits and drawbacks of virtual credit cards for:
- end users,
- the issuer (e.g., a bank).
## Part 2: Transaction validation rules
A transaction is described by:
- `card_number`: a 16-digit string
- `transaction_id`: an 8-digit string
- `amount`: integer dollars
Certain attributes are encoded in the digits:
### Transaction ID (8 digits, positions counted from left starting at 1)
- If the **7th** digit is `1` ⇒ transaction is **online**; if `0` ⇒ **offline/in-person**.
- If the **6th** digit is `1` ⇒ type is **charge**; if `0` ⇒ type is **authorization**.
### Card number (16 digits, positions counted from left starting at 1)
- If the **15th** digit is `1` ⇒ **merchant-bound**; if `0` ⇒ **not merchant-bound**.
- If the **14th** digit is `1` ⇒ network is **Master**; if `0` ⇒ **Visa**.
- If the **13th** digit is `1` ⇒ **multi-use**; if `0` ⇒ **one-time-use**.
### Validity rules by network
A transaction is valid if it satisfies the rules for its network:
**Visa**: valid iff
- (`merchant_bound` AND `multi_use`) OR
- (`online` AND `amount < 100` AND type is NOT `authorization`)
**Master**: valid iff
- (`amount < 100` AND (online OR offline) AND NOT `merchant_bound`) OR
- (`merchant_bound` AND `amount > 100`)
## Tasks
1. Describe an approach (data model + validation flow) to determine whether a transaction is valid.
2. For the following example, determine whether it is valid:
- `card_number = "1234567891011111"`
- `transaction_id = "50781100"`
- `amount = 150`
3. Explain how you would test this logic (unit tests and edge cases).
Quick Answer: This question evaluates understanding of virtual credit card product trade-offs and the ability to design data models and deterministic validation logic for encoded transaction attributes, testing competencies in payment systems, data encoding/decoding, and rule-based validation.
