美商AI Solution新星｜AI Infrastructure Engineer(PhD)｜博士在學 x 硬底投資 x 外商氛圍-Ting

Who Are We

A global AI computing company focused on delivering and measuring high-quality compute.

Role Overview

We are looking for elite talent at the intersection of Systems and Machine Learning (MLSys).

Requirements

This is not a standard "API-calling" machine learning role. It is a hardcore systems engineering and research position that requires a deep understanding of low-level GPU architecture, communication protocols, and memory management.

Key Responsibilities

• Distributed Computing Optimization: Design and optimize underlying architectures for distributed training and LLM inference serving across massive GPU clusters to break through performance bottlenecks.
• AI Infrastructure R&D: Contribute to the development of zSUITE and zFABRIC, improving fault tolerance, observability, and resource scheduling efficiency for large-scale computing clusters.
• Hardware-Software Vertical Integration: Collaborate with hardware engineering teams and top-tier OEM partners to fine-tune software performance for advanced architectures (e.g., advanced liquid cooling technologies, high-speed network topologies).
• Applied Frontier Technology: Track the latest academic breakthroughs in the MLSys domain (e.g., memory management, quantization algorithms, parallelization strategies) and translate them into practical commercial value for our Model-as-a-Service (MaaS) offerings.

Minimum Qualifications

• Ph.D. (or nearing completion) in Computer Science, Electrical Engineering, or a related field.
• Solid research background or hands-on experience in Distributed Systems, Parallel Computing, High-Performance Computing (HPC), or Deep Learning Systems.
• Proficiency in C++ and Python, with proven experience in CUDA programming or GPU performance profiling.
• Familiarity with mainstream distributed training and inference frameworks, such as DeepSpeed, Megatron-LM, vLLM, TensorRT-LLM, etc.
• Deep understanding of PyTorch internals and GPU memory management mechanisms.

Preferred Qualifications (Nice to Have)

• Publications in top-tier systems or machine learning conferences (e.g., OSDI, SOSP, NSDI, MLSys, ASPLOS, NeurIPS).
• Tuning experience with High-Performance Networking (RDMA, RoCE v2, InfiniBand) and collective communications (NCCL, MPI).
• Practical experience with large-scale cluster management and orchestration systems (e.g., Kubernetes, Slurm, Ray).
• Research or practical experience in Green AI, energy-aware computing, or GPU cluster power optimization.