CareerPlanSign in

Software Engineer, Workload Enablement

San Francisco💼 Full-time🗓 2026-03-28 → 2026-09-26

Core

Enabling production workloads and end-to-end testing on new AI platforms by porting inference and training systems, building stress benchmarks, and characterizing system behavior.

Role type

Senior IC software engineer (ML systems & performance)

Builds

Production-ready inference and training workloads, stress test suites, and performance characterization tools for new hardware platforms.

Domain

AI infrastructure, distributed systems, high-performance computing (HPC)

Deliverable

production ML models

Required skills

PyTorch, distributed training (data/model/pipeline parallel), RDMA, NCCL/RCCL, Python, C++, CUDA, performance profiling (Nsight, perf, flamegraphs), debugging

Preferred skills

workload-shaped benchmarking, RDMA networking tuning, Kubernetes integration, early hardware validation

Technologies

PyTorch, NCCL, RCCL, Nsight, rocprof, Kubernetes, RDMA, CUDA

Responsibilities

Port and validate inference/training workloads on new platforms; build E2E stress benchmarks; tune collective performance and overlap compute/communication; create repeatable CI test harnesses; partner on platform stability and scalability; produce bug reports and repros.

Seniority

Senior, hands-on IC

Sourced via ashby · Listed on CareerPlan, which tracks 70,000+ jobs from 20+ sources.