UMD / KMD - GPU Driver Development Expert

Position Summary

Samsung is a world leader in Memory, LCD and System LSI technologies. We are currently looking for exceptional software and hardware talent to join our Samsung Bangalore Office of Samsung’s strategic investments in high performance low power ARM based device technology. Presently our GPU design teams, located in Austin (SARC) , San Jose (ACL) & Bangalore SSIR , are developing a GPU that will be deployed in Samsung mobile products.

Role and Responsibilities

Design and document major units in a GPU pipeline targeted at Mobile graphics and machine learning

Develop functional and cycle simulators for GPU pipeline. Collect statistics to evaluate potential benefit of optimizations, prototype to test functional correctness and specify the scheme in detail for next level of implementation in hardware and or software

Collaborate with implementation, modeling, and software teams to define and develop microarchitecture, software implementation, and/or a verification plan

Investigate alternate approaches for important GPU workloads, incremental optimizations, and rebalancing to maximize performance in future key workloads

Work with software developers to understand important trends in future graphics and AI applications, problems faced by application and middleware developers

Find and/or implement applications to exercise novel algorithms in drivers/hardware

Minimum requirements: 4 to 18 Years

Experience using parallel programming APIs and languages such as OpenCL or CUDA

Knowledgeable in GPU or other parallel processing architectures , Strong knowledge of GPU architecture

Knowledge of APIs such as Vulkan, Open GS ES 3.x, DX 11/12, OpenCL and other compute APIs

Knowledge of key mobile GPU graphics workloads and compute application workloads such as computer vision, image processing, AI and Image compression

Knowledge of game applications, game rendering engines, academic papers on advanced rendering techniques desirable

Expertise in implementing advanced graphics rendering techniques, machine-learning (AI) approaches

Proven ability to debug complex issues in multi-threaded environments

Understanding of Operating System fundamentals and concepts

Familiarity with offline and JIT compiler designs

Background in Linux and Android development

Strong C, C++ and Python programming experience of industrial experience in systems programming (driver development a strong plus)

Strong algorithmic background and outstanding problem-solving skills

System level performance analysis and strong OS fundamentals (memory management, multithreading/synchronization, user/kernel mode interaction)

Excellent C and C++ programming skills (assembly a plus)

Familiarity with graphics APIs (e.g. Open GL/EGL, DirectX)

Understanding of rasterization pipeline and modern GPU architectures

Excellent communication and teamwork skills

Ability to own a problem and drive it to completion​

Preferred candidate will possess the following:

Experience working with ARM 64-bit architecture

Experience developing system software for Android OS

Knowledge of high-level shading languages, e.g., GLSL/HLSL

Understanding of modern real-time rendering game engines