Systems Performance Modeling Engineer
Core
Build simulation-based performance models and tools to predict how generative AI inference workloads perform on Tensordyne systems from single accelerators up to cluster scale.
Role type
Senior IC systems performance modeling engineer
Builds
Simulation models, trace-capture/replay tooling, and performance projections for multimodal generative AI inference systems
Domain
AI inference systems, custom silicon, hardware, networking, and distributed systems
Deliverable
production ML models
Required skills
C++, Python, distributed ML execution, collective communication, system architecture, performance modeling, simulation, experimental design, debugging
Preferred skills
LLM inference serving, collective communication libraries, data center/HPC networking, ML workload profiling, hardware/software co-design
Technologies
C++, Python, NCCL, RDMA, RoCE, GPUs, TPUs, ASICs
Responsibilities
Implement and extend simulation-based performance models for multimodal generative AI inference; Model serving strategies and their interaction with silicon and fabric topology; Build trace-capture and replay tooling; Model collective communication on multi-hop scale-out fabrics; Run calibration experiments and fix model errors; Run design-space sweeps and write analyses; Produce performance projections for product and customer discussions
Seniority
Senior, hands-on IC
