The AI Interview - Master AI/ML Interviews

Implement a basic agent execution loop that:

Starts with initial_input as state
Calls step_func(state) to get new state
Checks should_stop(new_state) for termination
Stops if should_stop returns True OR max_iterations reached
Returns final results

Safety Requirements:

Must respect max_iterations (hard stop)
Count iterations correctly
'terminated' is True if stopped by should_stop, False if by max_iterations
Handle case where should_stop is True immediately

Examples

Example 1:

Input: run_agent_loop(0, lambda x: x+1, lambda x: x>=5, 10)

Output: {'final_state': 5, 'iterations': 5, 'terminated': True}

Explanation: Increments from 0, stops at 5 (should_stop true), 5 iterations

Example 2:

Input: run_agent_loop(0, lambda x: x+1, lambda x: False, 3)

Output: {'final_state': 3, 'iterations': 3, 'terminated': False}

Explanation: Never meets stop condition, hits max_iterations, terminated=False

Starter Code

def run_agent_loop(initial_input, step_func, should_stop, max_iterations=10):
    """
    Execute a basic agent loop with termination conditions.
    
    Args:
        initial_input: Starting input for the agent
        step_func: Function(state) -> new_state
        should_stop: Function(state) -> bool (termination condition)
        max_iterations: Safety limit to prevent infinite loops
    
    Returns:
        dict with 'final_state', 'iterations', 'terminated' (bool)
    """
    # Your implementation here
    pass

Basic Agent Loop with Termination

Examples

Starter Code