Nvidia Solutions Architect

EY GDS Poland is part of the global integrated service delivery network by EY, operating across nine locations worldwide and collaborating with EY service lines, sectors, and geographies to deliver scalable AI solutions.

• 8+ years of experience in infrastructure and software engineering, including 2–3 years with NVIDIA technologies.
• Hands-on expertise with NVIDIA NIM, NeMo Agent Core, NeMo Guardrails, NeMo Retriever, TensorRT, and TensorRT-LLM.
• Strong skills in Python, C++, Linux, PyTorch, and OpenCV; experience with modern computer vision models such as YOLO, SAM, or VLMs.
• Experience with Kubernetes, Docker, NVIDIA GPU Operator, KServe, cloud platforms, and GPU environments such as DGX or InfiniBand.
• Experience with Kafka, GStreamer, RTSP streaming, and DeepStream SDK for real-time video analytics.
• Knowledge of LLMOps/MLOps, including model lifecycle governance and scalable deployment.
• Awareness of AI security, guardrails design, knowledge graphs, and graph-based retrieval.
• Experience with multimodal AI platforms, GPU capacity planning, and performance tuning.
• Background in regulated enterprise environments.
• Customer-facing or consulting experience supported by relevant NVIDIA or cloud certifications.

• Design and implement end-to-end AI architectures leveraging NVIDIA NIM, NeMo, DeepStream, and TensorRT.
• Build multi-agent architectures, agent orchestration patterns, and advanced RAG solutions using vector databases.
• Optimize inference performance and define GPU infrastructure.
• Collaborate cross-functionally to deliver robust, scalable, and production-grade AI solutions.
• Travel occasionally as required.

The position requires occasional travel. The company complies with the Whistleblower Protection Act and has established procedures for reporting misconduct via the EY Ethics Hotline.