Deep Learning Model and System Optimization Engineering Intern – 2027
Core
Optimizing inference performance, memory efficiency, and deployment cost for large language models on NVIDIA GPUs through hardware-software co-design.
Role type
Deep Learning Model and System Optimization Engineering Intern
Builds
GPU-accelerated software for deep learning inference, including inference frameworks like TensorRT-LLM
Domain
Artificial Intelligence / Large Language Models / GPU Computing
Deliverable
production ML models
Required skills
Transformer architectures, LLM inference frameworks, Python, PyTorch, CUDA C/C++, Triton, GPU architecture, low-precision computing, quantization, KV cache management, mixed-precision strategies
Preferred skills
Kernel optimization, speculative decoding, long-context inference, open-source contributions
Technologies
TensorRT-LLM, vLLM, SGLang, CUDA, Triton, NVIDIA GPUs, FP8, NVFP4, MXFP4
Responsibilities
Analyze and optimize inference performance across model architecture, numerical precision, GPU kernels, and memory systems; Develop low-bit quantization techniques for model weights and activations; Design and evaluate KV cache compression and sparsification techniques; Optimize model architectures including various attention mechanisms and state space models; Investigate model-system co-optimization techniques like speculative decoding; Build performance models for long-context inference and agentic workloads; Collaborate with model, framework, kernel, and GPU architecture teams
Seniority
Intern
