The AI Interview - Master AI/ML Interviews

Implement a knowledge graph memory system:

add_entity(id, type, attributes): Store entity with metadata
add_relation(subj, pred, obj, confidence): Store triple
query(entity_id, relation_type): Get entity and its relations
find_path(start, end, max_depth): BFS to find connection path

Entity Format: {'type': 'Person', 'attributes': {'name': 'Alice', 'age': 30}}

Relation Format: Tuple (subject_id, predicate, object_id, confidence)

Path Finding: Return list of predicates connecting start to end, or None if no path.

Examples

Example 1:

Input:

kg = KnowledgeGraphMemory(); kg.add_entity('a', 'Person', {'name':'Alice'}); kg.add_relation('a', 'knows', 'b', 0.9); 'knows' in [r[1] for r in kg.relations]

Output: True

Explanation: Relation added and retrievable

Starter Code

class KnowledgeGraphMemory:
    """
    Knowledge graph memory storing entities and relationships.
    Enables structured reasoning over connected facts.
    """
    
    def __init__(self):
        self.entities = {}  # id -> {'type': ..., 'attributes': {...}}
        self.relations = []  # (subject, predicate, object, confidence)
    
    def add_entity(self, entity_id, entity_type, attributes=None):
        """Add or update entity"""
        # Your implementation here
        pass
    
    def add_relation(self, subject, predicate, object, confidence=1.0):
        """Add relationship triple"""
        # Your implementation here
        pass
    
    def query(self, entity_id, relation_type=None):
        """
        Query graph: get entity info and related entities.
        If relation_type specified, filter by it.
        Returns {'entity': ..., 'relations': [...]}
        """
        # Your implementation here
        pass
    
    def find_path(self, start_id, end_id, max_depth=3):
        """Find path between two entities using BFS"""
        # Your implementation here
        pass

Simple Knowledge Graph Memory

Examples

Starter Code