The AI Interview - Master AI/ML Interviews

Implement a sliding window memory buffer that stores the most recent N messages. When the buffer is full and a new message arrives, the oldest message should be evicted (FIFO).

Requirements:

add(role, content): Add message as dict {'role': role, 'content': content}
get_recent(n): Return last n messages in chronological order
clear(): Remove all messages
Maintain max_size constraint strictly

Edge Cases:

n > current size should return all messages
n=None returns all messages
Empty memory should return empty list

Examples

Example 1:

Input: mem = SlidingWindowMemory(3); mem.add('user', 'Hi'); mem.add('assistant', 'Hello'); mem.get_recent()

Output: [{'role': 'user', 'content': 'Hi'}, {'role': 'assistant', 'content': 'Hello'}]

Explanation: Two messages added, both retrieved in order

Example 2:

Input: mem = SlidingWindowMemory(2); [mem.add('user', str(i)) for i in range(3)]; len(mem.messages)

Output: 2

Explanation: Max size 2, 3 messages added, oldest evicted

Starter Code

class SlidingWindowMemory:
    """
    Simple sliding window memory for conversation history.
    Maintains last N messages, evicting oldest when full.
    """
    
    def __init__(self, max_size=5):
        self.max_size = max_size
        self.messages = []
    
    def add(self, role, content):
        """Add message with role ('user', 'assistant', 'system')"""
        # Your implementation here
        pass
    
    def get_recent(self, n=None):
        """Get last n messages (or all if n is None)"""
        # Your implementation here
        pass
    
    def clear(self):
        """Clear all messages"""
        # Your implementation here
        pass

Sliding Window Memory Buffer

Examples

Starter Code