Uber-Style Event Routing Platform
Design the system architecture capable of handling millions of concurrent rider and driver location updates without dropping messages.
┌─────────────┐ ┌──────────────┐ ┌──────────────┐
│ SOURCES │ │ INGESTION │ │ PROCESSING │
│ │ │ │ │ │
│ Rider App │────▸│ Kafka Topics │────▸│ Flink Jobs │
│ Driver App │ │ 50 topics │ │ Windowing │
│ Payments │ │ 1B events/d │ │ Enrichment │
│ Maps API │ │ │ │ │
└─────────────┘ └──────────────┘ └──────┬───────┘
│
┌─────────────────────────┘
│
▼
┌─────────────┐ ┌──────────────┐ ┌──────────────┐
│ SERVING │ │ STORAGE │ │ GOVERNANCE │
│ │ │ │ │ │
│ Driver ETA │◂────│ Iceberg │ │ SLA/SLOs │
│ Surge Price │ │ Hot/Warm/Cold│ │ Observability│
│ Analytics │ │ 5 PB total │ │ Cost Control │
│ ML Features │ │ │ │ │
└─────────────┘ └──────────────┘ └──────────────┘Fig 1.1: Uber event platform architecture -- 6 layers, 1B events/day
What You'll Design
Architecture Blueprint
Six-layer platform architecture handling 1B events/day across ride, payment, and mapping domains.
Scale Engineering
Capacity planning from startup (10K events) to hyperscale (1B events/day) with cost modeling at each tier.
SLA & Reliability
99.99% availability SLAs with disaster recovery, backfill strategy, and incident response playbooks.
Cost Tradeoffs
Compute vs storage vs latency optimization. Justify every dollar of your $2.4M/year platform budget.
Progressive Build Path
Each part builds on the previous. Watch your platform design evolve from requirements to a complete staff-level architecture.
Your Design Challenge
You are the Staff Data Engineer at Uber. The VP of Data asks you to redesign the event platform.
THE BRIEF
You've just been promoted to Staff Engineer. Your first project: redesign Uber's event platform from scratch.
CONSTRAINTS
- •4 source systems: Rider App, Driver App, Payments, Maps
- •Current: 10K events/day (startup phase)
- •Target: 1B events/day within 18 months
- •Budget: Start at $50K/month, justify every increase
- •SLA: 99.99% availability, <5 min latency for analytics
- •Compliance: GDPR (EU riders), CCPA (US riders)
- •Team: 4 data engineers (you're hiring 2 more)
DELIVERABLE
A complete system design document with architecture diagrams, capacity estimates, cost projections, and trade-off rationale.
What Makes This Different
No Code, All Architecture
Pure whiteboard exercises. You design systems using diagrams, trade-off matrices, and capacity calculations -- exactly like a Staff interview.
Constraints-Driven
Every decision has a cost. Budget caps, SLA requirements, and team size force real trade-offs -- not academic ideals.
Progressive Complexity
Start at 10K events/day. By Part 4, you're defending a 1B events/day platform with disaster recovery and $2.4M annual budget.
Prerequisites
- Familiarity with data platform concepts (Kafka, Spark, data lakes)
- Understanding of distributed systems basics
- 2+ years of data engineering experience (recommended)
- Complete System Design skill toolkit first (or have equivalent knowledge)
Related Skill Toolkit
This project pairs perfectly with the System Design for Data Engineers skill toolkit. Complete the modules first for maximum understanding, or dive straight in if you have prior system design experience.
View System Design Skill ToolkitNew to Kafka? Read the complete guide to topics, partitions, consumer groups, and retention before diving in.
What is Apache Kafka? — Full GuideReady to design?