Isaiah Witzke developed and integrated a 50/50 A/B testing framework for default LLM model selection in the warpdotdev/warp repository, targeting free-tier user onboarding. By leveraging Rust and experimentation best practices, Isaiah enabled the assignment of users to either an “auto-open” or “auto-cost-efficient” model, collecting telemetry to inform future default choices. This work established a foundation for data-driven decision making in onboarding workflows, balancing performance and cost considerations. The implementation required careful integration with existing telemetry systems and ensured stability throughout the experiment, demonstrating Isaiah’s proficiency in A/B testing, LLM integration, and user onboarding within a production Rust codebase.