Break the Iceberg (Order)

Detecting iceberg orders in high-frequency trading interview

This hard Python coding problem tests your ability to model and detect hidden order structures in a live market feed. It combines graph construction, time-windowed pattern matching, and order-matching logic—all skills that matter in market microstructure roles at quant trading firms.

The core challenge is to recognize when a sequence of orders, submitted within a one-minute window and sharing price, quantity, and side, forms a chain of executions consistent with an iceberg strategy. You must track which orders are fully filled, match incoming orders against the book using price-time priority, and then build a directed graph where edges represent the parent-child relationships between suspected iceberg tranches. The root nodes—those with no parents—become your output.

• Order-book state management and FIFO matching rules
• Timestamp arithmetic and time-windowed event detection
• Graph construction from temporal and structural constraints
• Edge cases: multiple parents, multiple children, partial fills, and competing orders at the same price