Design an A/B testing platform

Design an internal A/B testing platform for a consumer product company. Product and engineering teams should be able to: - create, update, pause, and archive experiments - define control and treatment variants - target experiments to specific user segments - assign users consistently to variants across sessions - log exposure, click, conversion, and custom events - define primary metrics and guardrail metrics - monitor experiment results in dashboards - detect obvious data quality issues such as sample ratio mismatch Your design should cover: - APIs and core data model - experiment configuration storage - user bucketing and sticky assignment - client/server SDKs for logging exposure and events - event ingestion and storage - metric computation and statistical analysis - reliability, scalability, and low-latency requirements - common pitfalls such as double counting, event loss, and invalid experiment reads