Axon·Software Engineer·Onsite - System Design / Architecture·Senior

SeniorPrefer not to say

Apr 2026

Summary

Axon system design round focused entirely on building a reporting pipeline across microservices. Pretty deep dive, way more nuanced than I expected for what I thought would be a standard design question.

Questions Asked (1)

Design a reporting system for a microservice architecture that aggregates metrics, business events, and operational data to support ad-hoc queries and dashboards. Walk through how you'd handle data collection, storage, transformations, query performance, and the trade-off between data freshness and cost.

System DesignTechnical Trade-offsData Modeling

Author's notes

This question is way broader than it sounds.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying requirements: data sources, query patterns, freshness needs, and budget constraints. Then propose a layered architecture: ingestion (streaming + batch), storage (data lake + warehouse + OLAP), transformation (ETL/ELT), and serving (query engine + dashboards). Finally, discuss trade-offs between freshness and cost, and how to optimize query performance.

Pro tip: Emphasize that you'd start with a simple, cost-effective solution (e.g., batch ETL into a warehouse) and only add streaming complexity when freshness requirements justify the cost. This shows pragmatism and business awareness.

1. Clarify Requirements

Ask about data volume, variety, velocity, query latency, freshness SLAs, and budget. Identify key stakeholders and use cases (ad-hoc vs. dashboards).

2. Design Data Collection

Propose a unified ingestion layer: use Kafka for real-time events and metrics, and batch ingestion for historical data. Ensure schema management and data quality checks.

3. Choose Storage & Transformation

Store raw data in a data lake (S3) for flexibility, and transformed data in a warehouse (Snowflake/BigQuery) or OLAP cube (Druid/ClickHouse) for fast queries. Use ETL/ELT pipelines (Airflow, dbt) to clean and aggregate.

4. Optimize Query Performance

Implement partitioning, indexing, and pre-aggregation. Use a query engine that supports caching and concurrency. Consider materialized views for common dashboard queries.

5. Balance Freshness vs. Cost

Offer tiered freshness: real-time for critical metrics (streaming), micro-batch for near-real-time, and daily batch for historical. Discuss cost implications of each tier and how to monitor and adjust.

Key Points to Mention

Lambda vs. Kappa architecture for handling batch and stream processing
Use of columnar storage and compression for cost-effective analytics
Data modeling: star schema vs. wide tables for ad-hoc queries
Query performance techniques: indexing, partitioning, caching, and pre-aggregation
Trade-offs: freshness (latency) vs. cost (compute, storage, complexity)
Monitoring and alerting on data pipelines and query performance

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.