Staff Software Engineer, (DEM)
Crusoe · San Francisco, CA - US
About this role
## About Crusoe Crusoe is on a mission to accelerate the abundance of energy and intelligence. As a vertically integrated AI infrastructure company, we own and operate each layer of the stack—from electrons to tokens—to power the world’s most ambitious AI workloads. ## Role: Staff Software Engineer (DEM) Power, not silicon, is the binding constraint on AI infrastructure. Crusoe’s **Dynamic Power Management (DPM)** system measures, forecasts, and orchestrates energy use across our data centers—from the site level down to the individual GPU—so power goes where it does the most work. This is a **mission-critical, ground-up** system that runs underneath every customer workload. As a founding engineer, you’ll partner with the principal architect to turn product requirements into a production architecture, ship the first version, and help shape the team that grows around it. ### What you’ll be working on - **DPM control loop:** Design and build the system that allocates power across site/row/rack/node/GPU based on real demand, applies decisions through GPU/chassis power management, and adapts within seconds. - **Power telemetry:** Own the pipeline for high-frequency power data from GPUs, NICs, chassis/BMCs (Redfish/IPMI), PDUs, and facility systems—handling latency, accuracy, and stale/missing data like a control system requires. - **Power modeling & forecasting:** Determine how much power each site/row/rack truly needs, identify idle capacity, and quantify confidence using live telemetry + simulation. - **Workload energy profiling:** Understand how real training/inference workloads consume energy and translate findings into efficiency policies. - **Simulation & trace replay:** Build tools to replay production power/workload traces so policy changes can be validated offline. - **Shadow-to-production rollout:** Run DPM in shadow mode first, compare recommendations to actual usage, and automate only once the system earns trust. - **Safety & failure design:** Define fail-safe defaults, behavior on stale telemetry, limits on adjustment rates, audit trails, and verification of intended effects. - **Capacity decisions:** Connect DPM to capacity planning and VM placement. - **Metrics & tradeoffs:** Define success measures (e.g., compute delivered per megawatt, performance per watt) and communicate gains/costs across engineering, energy, finance, and leadership. - **Cross-functional leadership:** Partner with data center electrical/mechanical engineering, the energy team, capacity planning, and hardware vendors (e.g., NVIDIA). Lead design reviews and mentor engineers. ### What you’ll bring - **10+ years** experience (or **8+ with clear technical leadership**) building and operating distributed production systems; experience taking a system from design through launch. - Experience building **control systems / closed-loop automation** where incorrect outputs have physical or financial consequences (e.g., autoscalers, admission control, rate limiting, power/thermal management, industrial control software). - Ability to turn forecasting/simulation into **production services** other systems depend on. - Strong depth in **high-volume time-series telemetry** (ingestion, aggregation, staleness/gaps/measurement error). - Fluency in **Go or Rust**. - Comfort working with hardware and facility teams; ability to read an electrical one-line diagram (or learn quickly). - Ability to explain technical risk in **megawatts, dollars, and customer impact**, and adju
Listing freshness
CronJobs last confirmed this listing 2h ago. If its source stops confirming the opening for seven days, this page is removed from active inventory.