Neural Rendering Research Inference Engineer – Advanced Graphics Programs

Nasze wymagania: Strong object-oriented programming background, C/C++ preferred Proven experience in developing and optimizing GPU kernels for machine learning workloads (eg:, using HIP, OpenCL, CUDA, HLSL ) Ability to program in low level languages (x86 asm, SSE, ISA, Ptx, AMD assembly) Strong understanding of GPU architecture (compute cores, cache hierarchy, memory model), graphics APIs (DirectX, OpenGL, Vulkan, etc.), and shader programming. Solid understanding of common neural network operators, their mathematical foundations, and computational requirements. Experience with modern concurrent programming and threading APIs Experience with Windows, Linux operating system development Experience with software development processes and tools such as debuggers, source code control systems (GitHub) and profilers is a plus Understanding Machine Learning techniques and its application within graphics Effective communication and problem-solving skills Good interpersonal skills Undergrad degree required. Master's degree or PhD in Computer Science especially Graphics and/or Machine Learning, Computer Engineering, or equivalent preferred. Zakres obowiązków: Work with ML researchers and engineers to translate neural network models and algorithms written in PyTorch/Onnx to efficient GPU shaders using languages such as HIP, Cuda, HLSL Design, implement and optimize high performance GPU kernels for ML operators. Work across research, hardware, driver and compiler teams to analyze and troubleshoot performance issues, provide solutions to improve rendering speed and ML workload efficiency. Stay current with latest advancements in GPU hardware, rendering techniques, graphics APIs and GPU accelerated ML. Contribute to the design and development of tools and methodologies for optimized shader integration to game engines. Document and share knowledge on best practices for GPU programming (both graphics and compute/ML) within the team. Participate in code reviews and provide constructive feedback to peers.