Performance Engineer

Advanced Micro Devices, Inc•San Francisco, CA

About The Position

The Performance Engineer will help make World Labs’ generative world models train and serve as efficiently as the hardware allows. This hands-on individual-contributor role focuses on identifying and eliminating bottlenecks across GPU kernels, inference and serving paths, training loops, and GPU utilization. The engineer will work across the stack, from low-level tensor and kernel optimization to fleet-wide serving efficiency, partnering closely with researchers to accelerate their models. The role’s impact is measured through improved throughput, reduced latency, increased utilization, and preserved correctness.

Requirements

  • Strong performance-engineering foundations, including profiling, roofline analysis, latency/throughput optimization, and disciplined root-cause investigation.
  • Deep GPU programming and optimization experience in CUDA and/or Triton, including kernel tuning, memory hierarchy, and bandwidth optimization at scale.
  • Hands-on experience optimizing inference and serving for large models, including batching, KV/prompt caching, quantization, and low-latency, high-throughput sampling.
  • Hands-on experience optimizing training performance, including parallelism, distributed communication, mixed/low precision, and utilization.
  • Working knowledge of ML framework internals such as PyTorch and/or JAX, and compiler paths such as torch.compile or XLA.
  • Strong proficiency in Python and ability to work in C++/CUDA, and Rust or Go as needed.
  • High-ownership approach focused on shipped throughput improvements and latency reductions.
  • BS, MS, or PhD in Electrical Engineering, Computer Science, Computer Engineering, or a closely related field.

Nice To Haves

  • Experience at an AI lab or ML-native company, optimizing systems used directly by researchers and productionizing research code.
  • Depth in low-precision numerics, including FP8/INT8 quantization, mixed precision, and detecting numerical regressions across hardware platforms.
  • Distributed-systems experience for large-scale training and inference, including NCCL, NVLink, model and tensor parallelism, and fault tolerance.
  • Experience serving generative, diffusion, video, or 3D/spatial models, beyond text LLMs.
  • Experience with multiple accelerator types, such as GPU plus TPU or Trainium, and partnering with hardware vendors on accelerator capabilities.
  • Experience building performance-modeling and observability frameworks for GPU utilization and cost.

Responsibilities

  • Optimize inference and serving end to end, including latency, throughput, batching, caching, and scheduling, for production-scale models.
  • Write and tune GPU kernels using CUDA and/or Triton, including kernel fusion, memory- and bandwidth-bound optimization, and low-precision execution.
  • Improve training throughput and GPU utilization through parallelism strategies, communication/compute overlap, mixed precision, and elimination of pipeline stalls.
  • Build performance models, profiling workflows, and observability to make throughput, latency, cost, utilization, and tradeoffs visible across the stack.
  • Maintain numerical correctness across precision, kernel, and hardware changes.
  • Partner with researchers to productionize models for serving and make experiments faster and more reliable.
  • Where needed, work on distributed systems used for training and inference, while focusing primarily on maximizing GPU performance.

Benefits

  • AMD benefits are referenced as “AMD benefits at a glance,” but specific benefits are not listed in the posting.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service