Automated A/B Experimentation Platform

You are an experimentation platform engineer building automated A/B testing infrastructure for games: Game type: [INSERT - mobile, live service, multiplayer] Experiment volume: [INSERT - concurrent tests] Decision speed: [INSERT - hours to weeks] Stack: [INSERT - client config, server feature flags, analytics warehouse] Cover the full pipeline: 1. Design - experimentation culture and hypothesis templates, guardrail vs success metrics, OEC definition, who can launch and approve tests. 2. Assignment - stable user/device hashing, layered and mutually exclusive experiments, targeting by segment/region, cross-device identity, sticky assignments. 3. Technology - feature flag and config service, SDKs for client and server, real-time exposure logging, kill switches, bandwidth-safe config sync on mobile. 4. Statistics - power analysis and sample sizing, sequential and Bayesian options, novelty effects, variance reduction, ratio and distribution metrics like spend. 5. Data - exposure and event joins, automated dashboards, guardrail alerts (crash, latency, refund), data freshness and missing-event monitoring. 6. Testing - A/A tests, assignment-balancing checks, leakage and interaction tests between layers, telemetry correctness validation. 7. Launch - peer review and ship gates, rollback plans, compliance and consent handling, regional pilot before global.