Intern, AI Application Engineer
Core
Intern role deploying and tuning LLM inference serving on data-center GPUs and building automated benchmark harnesses for storage architecture.
Role type
Intern, AI Application Engineer (LLM Inference & Storage)
Builds
Automated benchmark harnesses for LLM inference performance and storage requirements analysis
Domain
AI/ML Infrastructure, Data Center Storage, GPU Systems
Deliverable
production ML models
Required skills
LLM inference serving (vLLM, SGLang, Hugging Face TGI), Python, PyTorch, Linux, NVIDIA GPUs, KV cache management, latency/throughput benchmarking, RDMA/NVMe-oF familiarity, containerization (Docker/Kubernetes)
Preferred skills
Master's/PhD in CS/CE/EE, understanding of LLM serving internals (prefix caching, parallelism), FPGA/SoC collaboration experience
Technologies
vLLM, LMCache, NVIDIA Dynamo/NIXL, NVIDIA H200, BlueField-3, NVMe-oF, RDMA, RoCE, Python, PyTorch, Docker, Kubernetes, Git
Responsibilities
Deploy and operate LLM inference service on 4x NVIDIA H200 GPU server, Build automated benchmark harness for long-context/agent-style traces, Configure KV cache offloading to network-attached storage, Run daily experiment sweeps and analyze bottlenecks, Collaborate with FPGA/SoC/firmware engineers to debug data path, Summarize storage requirements for future SSD controllers
Seniority
Intern