Intercom - Product Engineer - II Experience (Dublin)
- Location
- Dublin
- Posted
- 25 Jun 2025
Summary
This post details a coding interview problem from Intercom where I was tasked with designing and implementing an AssignmentSystem to distribute customer conversations among support agents efficiently, considering individual limits and load balancing rules.
How it went
In this exercise, you'll build a system to assign conversations to customer support agents.
Task
Implement an AssignmentSystem class that has the following API:
- Initializes with a list of available agents.
set_limit(agent_name, limit)- Sets the conversation limit for a specific agent.
assign(conversation_id)- Assigns a conversation to the next available agent.
preview_assignments(count)- Returns a list of agents who would be assigned conversations next.
Requirements
- When assigning conversations, balance load evenly:
- Assign new conversations to the agent with the fewest conversations
- If there are ties, pick the agent who's been waiting the longest since their last assignment
- Each agent has a maximum conversation limit (default is 2)
Example:
agents = ["Alice", "Bob", "Charlie"]
system = AssignmentSystem(agents)
system.set_limit("Bob", 4)
system.set_limit("Charlie", 3)
I want to know who will receive next 4 conversations
system.preview_assignments(4)
Output: ["Alice", "Bob", "Charlie", "Alice"]
Make some assignments
system.assign(101) # Assigns to Alice
system.assign(102) # Assigns to Bob
system.assign(103) # Assigns to Charlie
system.assign(104) # Assigns to Alice
I want to know who will receive next 5 conversations
system.preview_assignments(5)
['Bob', 'Charlie', 'Bob', 'Charlie', 'Bob']
Questions asked (1)
- 01
Conversation Assignment System
Data Structures & Algorithms · Medium
In this exercise, you'll build a system to assign conversations to customer support agents.
Task
Implement an AssignmentSystem class that has the following API:
- Initializes with a list of available agents.
set_limit(agent_name, limit)- Sets the conversation limit for a specific agent.
assign(conversation_id)- Assigns a conversation to the next available agent.
preview_assignments(count)- Returns a list of agents who would be assigned conversations next.
Requirements
- When assigning conversations, balance load evenly:
- Assign new conversations to the agent with the fewest conversations
- If there are ties, pick the agent who's been waiting the longest since their last assignment
- Each agent has a maximum conversation limit (default is 2)
Example:
agents = ["Alice", "Bob", "Charlie"]
system = AssignmentSystem(agents)
system.set_limit("Bob", 4)
system.set_limit("Charlie", 3)
I want to know who will receive next 4 conversations
system.preview_assignments(4)
Output: ["Alice", "Bob", "Charlie", "Alice"]
Make some assignments
system.assign(101) # Assigns to Alice
system.assign(102) # Assigns to Bob
system.assign(103) # Assigns to Charlie
system.assign(104) # Assigns to Alice
I want to know who will receive next 5 conversations
system.preview_assignments(5)
['Bob', 'Charlie', 'Bob', 'Charlie', 'Bob']