The AI Interview - Master AI/ML Interviews

Implement agent lifecycle state management:

States: IDLE, RUNNING, PAUSED, STOPPED, DONE, ERROR

Valid Transitions:

IDLE → RUNNING (start)
RUNNING → PAUSED (pause), STOPPED (abort), DONE (complete), ERROR (exception)
PAUSED → RUNNING (resume), STOPPED (abort)
ERROR is terminal (no exits)
STOPPED is terminal
DONE is terminal

Requirements:

transition_to(new_state): Returns True if valid, False if invalid
Record all transitions in state_history as {'from': X, 'to': Y}
can_execute(): Only True in RUNNING state

Examples

Example 1:

Input: life = AgentLifecycle(); life.transition_to('RUNNING'); life.state

Output: 'RUNNING'

Explanation: Valid transition from IDLE to RUNNING

Example 2:

Input: life = AgentLifecycle(); life.transition_to('DONE')

Output: False

Explanation: Cannot go from IDLE directly to DONE

Starter Code

class AgentLifecycle:
    """
    Manage agent lifecycle states: IDLE → RUNNING → PAUSED/STOPPED → DONE
    """
    
    VALID_STATES = ['IDLE', 'RUNNING', 'PAUSED', 'STOPPED', 'DONE', 'ERROR']
    
    def __init__(self):
        self.state = 'IDLE'
        self.state_history = []
    
    def transition_to(self, new_state):
        """
        Transition to new state if valid.
        Valid transitions:
        IDLE → RUNNING
        RUNNING → PAUSED, STOPPED, DONE, ERROR
        PAUSED → RUNNING, STOPPED
        ERROR → (no transitions)
        Returns True if successful, False otherwise
        """
        # Your implementation here
        pass
    
    def can_execute(self):
        """Check if agent can execute actions (only in RUNNING)"""
        # Your implementation here
        pass

Agent State Machine

Examples

Starter Code