Staff Engineer – Experimentation Team

<h4><strong>About the Job:</strong></h4> <p data-renderer-start-pos="5871" data-local-id="4e17a2683c0d">As a Staff Engineer on LaunchDarkly's Experimentation team, you'll build the platform that helps engineering teams make data-driven decisions with confidence. Our Experimentation product enables customers to run A/B tests, measure the impact of feature changes, and optimize experiences — integrated with a feature management platform that processes trillions of evaluations daily.</p> <p data-renderer-start-pos="6254" data-local-id="ca91cb8d3f76">This role sits at the intersection of data science and platform engineering. You'll design the statistical engine, warehouse-native analysis pipelines, and adaptive experimentation systems (including contextual bandits) that power our customers' most important decisions. We want someone who brings genuine depth in applied statistics and ML — as fluent in statistical validity as in system architecture.</p> <p data-renderer-start-pos="6660" data-local-id="b9420194823c">You'll also architect warehouse-agnostic features that run analysis directly inside customers' data warehouses (Snowflake, Databricks, Redshift, BigQuery) — modular computation layers that abstract across warehouse environments while maintaining statistical correctness.</p> <p data-renderer-start-pos="6932" data-local-id="48ac074b511d">Deep technical experience, a scientific mindset, and the ability to influence product and technical direction are critical. You'll lead by example: setting the bar for rigor, mentoring teammates, and owning systems end to end, including on-call.</p> <h4>Responsibilities:</h4> <ul class="ak-ul" data-local-id="f0d41202-4500-43a9-8b4b-6ddfd7404ab8" data-indent-level="1"> <li> <p data-renderer-start-pos="7204" data-local-id="1d99782c9741">Build the experimentation statistical engine — hypothesis testing, sequential analysis, variance reduction (<span data-highlighted="true" data-vc="highlighted-text">CUPED</span>, Winsorization), power analysis. Ensure statistical correctness across all experiment types.</p> </li> </ul> <ul class="ak-ul" data-local-id="c487b615106e" data-indent-level="1"> <li> <p data-renderer-start-pos="7416" data-local-id="a37de6d4875e">Design warehouse-native experimentation that runs analysis inside customer warehouses (Snowflake, Databricks, Redshift, BigQuery). Build modular, warehouse-agnostic abstractions for rapid new backend support.</p> </li> </ul> <ul class="ak-ul" data-local-id="4312e12305d6" data-indent-level="1"> <li> <p data-renderer-start-pos="7630" data-local-id="88e21a3c73d8">Lead adaptive experimentation — contextual bandit systems, Bayesian optimization, automated allocation beyond simple A/B tests.</p> </li> </ul> <ul class="ak-ul" data-local-id="bc1cee49accd" data-indent-level="1"> <li> <p data-renderer-start-pos="7763" data-local-id="1375252d5b86">Drive the platform roadmap with product, design, and data science. Shape what we build, not just how.</p> </li> </ul> <ul class="ak-ul" data-local-id="7b7a37e69988" data-indent-level="1"> <li> <p data-renderer-start-pos="7870" data-local-id="77443be3a78a">Collaborate cross-functionally with Warehouse Integrations, SDK, Platform, and Data Science teams.</p> </li> </ul> <ul class="ak-ul" data-local-id="084267b0c466" data-indent-level="1"> <li> <p data-renderer-start-pos="7974" data-local-id="ea78298a8ace">Mentor engineers and raise the team's bar for statistical rigor and system design.</p> </li> <li> <p data-renderer-start-pos="8062" data-local-id="66703d816155">Own operational excellence — monitoring, observability, incident response, on-call. Robust telemetry and alerting.</p> </li> </ul> <h4>Qualifications:</h4> <ul class="ak-ul" data-local-id="3d814296-4fde-47e7-b0ca-34d2aae9ecc1" data-indent-level="1"> <li> <p data-renderer-start-pos="8201" data-local-id="8059a191b42d">10+ years building large-scale experimentation platforms, statistical analysis systems, or data-intensive backend services.</p> </li> </ul> <ul class="ak-ul" data-local-id="418c4277f9af" data-indent-level="1"> <li&g