Senior Data Scientist, Engine Infra
Roblox, San Mateo, CA
What you will do
- Define Metrics – Own Engine Infra metrics, telemetry logging, and validation for the RCC fleet and edge compute. Establish KPIs for game-server efficiency, reliability, and player experience — including CPU and memory, heartbeat, occupancy and multiplexing efficiency, crash and OOM rates, runtime error codes, and latency quality (e.g. DataPing vs RakPing) that directly affect creators and players.
- Transform & Visualize Data for Actionable Insights — Access and process raw RCC and engine telemetry, build analytics pipelines, and deliver clear operating views of the fleet. Turn per-instance, per-game, datacenter, and channel signals into dashboards and decision systems that help Engine Infra pack more players onto fewer cores, allocate CPU and memory correctly, and keep multiplexed and large-server processes healthy.
- Design Experiments & Apply Causal Inference — Lead channel tests, hardware A/Bs, and causal analyses for multiplexing, ARM64 bursting, regional capacity, and edge latency. Measure trade-offs among efficiency, stability, and quality; separate host, container, engine, content, and release effects; and estimate player-experience impact (including latency-sensitive cohorts, not only global averages) so launches ship on evidence.