AI Performance Engineer, Cloud AI (Multiple Levels: Entry-Level to Senior Staff) (Hsinchu/Taipei)
Core
Optimizing and deploying large language models (LLMs), vision-language models (VLMs), and diffusion models for efficient inference on Qualcomm hardware.
Role type
Senior IC AI Performance Engineer (Cloud AI)
Builds
Inference acceleration solutions for GenAI workloads
Domain
Cloud AI / Machine Learning / Semiconductor Hardware
Deliverable
production ML models
Required skills
PyTorch, ONNX, transformer architectures, attention mechanisms, Python, Triton, workload mapping, sharding, parallelisms, computer architecture, ML accelerators, distributed systems
Preferred skills
neural network operators, machine learning compilers, torch.compile, torchDynamo, accuracy evaluation methods
Technologies
PyTorch, ONNX, Triton
Responsibilities
Convert, optimize and deploy models for efficient inference; Analyze complex performance or stability issues to identify root causes; Design and implement high-level kernels in Triton; Map next generation AI workloads on hardware designs; Collaborate with compiler, firmware and platform teams; Scale performance for throughput and latency constraints
Seniority
Senior, hands-on IC
