README.md

This is the starter kit for the problem statement defined for interface.ai's Data Engineer role

Problem Statement

Multi-tenant Windowed Aggregation

Your team runs a multi-tenant data platform that receives a stream of transaction events from different customers (tenants). Each event has the following fields:

{
    "tenant_id": str,
    "user_id": str,
    "event_time": str, # ISO-8601 in UTC, e.g. "2025-01-01T10:03:21Z"
    "amount": float,
}

You are given an iterable of events. You need to group events into tumbling time windows and compute per-tenant aggregates.

Requirements:

1. Use tumbling time windows of fixed length window_minutes in UTC.
2. For each event:

• Parse event_time as a UTC timestamp.
• Assign the event to a window based on its timestamp.

3. Windows are half-open intervals: window_start <= event_time < window_end Example with 15-minute windows:

• 10:00:00Z – 10:15:00Z
• 10:15:00Z – 10:30:00Z
• 10:30:00Z – 10:45:00Z

4. Window calculation:

• window_start is truncated down to the nearest multiple of window_minutes since the Unix epoch.
• window_end = window_start + window_minutes

5. Aggregate per (tenant_id, window). For each combination, compute:

• tenant_id: str
• window_start: ISO-8601 UTC string, e.g. "2025-01-01T10:00:00Z"
• window_end: ISO-8601 UTC string
• event_count: number of events
• unique_users: number of distinct user_id values
• total_amount: sum of amount
• avg_amount_per_event: total_amount / event_count (or 0.0 if no events)

6. Return a list of dictionaries with exactly these keys:

   {
       "tenant_id": str,
       "window_start": str,
       "window_end": str,
       "event_count": int,
       "unique_users": int,
       "total_amount": float,
       "avg_amount_per_event": float,
   }
   

7. Sort output by: tenant_id ascending, then window_start ascending. Assumptions:

• event_time is always valid in format "%Y-%m-%dT%H:%M:%SZ"
• amount can be safely cast to float

How to execute

./run.sh "[{\"tenant_id\": \"t1\", \"user_id\": \"u1\", \"event_time\": \"2025-05-10T12:34:56Z\", \"amount\": 100.0}]"