Resolving Technical Disagreements

Level: 201 (Senior Staff)
Category: Working with Others & Collaboration

Question: Tell me about a time when you had a significant technical disagreement with another engineer. How did you resolve it and ensure the team moved forward effectively?

Strong Answer Example

As senior engineer at Airbnb, I had a fundamental disagreement with Alex, a senior colleague, about our approach to rebuilding the search infrastructure. I advocated for an event-driven architecture with eventual consistency, while Alex insisted on a synchronous approach with strong consistency guarantees.

The Context:
The disagreement had been ongoing for 3 weeks, causing delays in our quarterly planning. Both approaches had merit, but we were at an impasse, and the team was starting to take sides.

My Resolution Approach:

1. Create Shared Context:
I acknowledged that both approaches had valid concerns:
"My approach prioritizes scalability and resilience, but Alex's approach prioritizes data consistency. Let's define what we actually need to achieve."

2. Establish Decision Criteria:
Rather than arguing for our individual preferences, I proposed we define shared success metrics:

• Search response time under 200ms
• 99.9% uptime during peak periods
• Ability to handle 10x current search volume
• Data consistency within acceptable limits for search results

3. Run Technical Experiments:
I suggested we prototype both approaches in a controlled environment:

• Built a proof-of-concept of Alex's synchronous approach
• Built a proof-of-concept of my event-driven approach
• Measured performance, complexity, and operational overhead for each

4. Data-Driven Discussion:
We presented results to the broader team:

• Synchronous approach: 150ms response time, 99.7% uptime, complex deployment
• Event-driven approach: 120ms response time, 99.95% uptime, simpler operations
• Both met consistency requirements for 95% of use cases

5. Collaborative Decision:
Based on the data, we chose a hybrid approach: event-driven core with synchronous fallback for critical operations. Alex contributed the consistency monitoring framework, while I designed the event processing pipeline.

Results:
The hybrid solution was 30% more performant than either original approach, and both Alex and I became strong advocates for data-driven technical decisions.