NVIDIA GPU Engineer for AI Deployment

We’re looking for an experienced engineer to set up and optimise a local AI environment running on NVIDIA hardware. We need someone who can handle deployment, troubleshoot GPU issues, and deliver a working setup with clear documentation. The work includes: Configuring NVIDIA drivers, CUDA, and the required runtime. Deploying suitable open-source LLMs for local inference. Optimising GPU memory usage, response speed, and concurrent requests. Setting up an inference API for integration with our application. Testing performance and documenting installation, maintenance, and recovery steps. You should have hands-on experience with NVIDIA GPUs, Linux, Docker, and local LLM deployment. Experience with inference tools such as TensorRT-LLM, vLLM, or llama.cpp is relevant, depending on hardware compatibility. The first milestone will be to assess our hardware and workload, recommend a suitable deployment approach, and get one model running with measured performance. Please answer these questions when applying: Which NVIDIA hardware have you used for production AI workloads? Describe a similar deployment you completed and your responsibilities. How would you measure and improve inference performance? What information would you need to estimate the timeline and cost?