Senior GPU Compiler Engineer – Parallel Computing

Job Summary

Do you want to work at the intersection of compiler writing and CUDA GPU development?  We are seeking a talented, self-motivated and self-directed senior compiler engineer with good working knowledge of CUDA GPU development.  You will be responsible for expanding and improving our dynamic compilation of MATLAB onto parallel hardware platforms such as GPUs.  This is a challenging problem that requires an understanding of compiling a dynamic language such as MATLAB and complicated target platforms.

This will require excellent object-oriented design and programming skills, strong inter-personal and communication skills, and excellent compiler and GPU implementation experience.  The team you will be joining is responsible for the Parallel Computing Toolbox and the MATLAB Distributed Computing Server.  

Responsibilities

• Contribute to new code generation pipeline targeting GPUs
• Design and implement wide-ranging performance optimizations
• Make high performance computing architectures available in MATLAB
• Work with teams focused on MATLAB execution engines for heterogeneous, multi-threaded environments
• Participate in all team activities such as planning, design discussions and reviews

Qualifications Required / Preferred

• BS in Computer Science, Engineering ,Math, Physics, + 7 years experience or MS, PhD or equivalent or a related field and 3-5 years relevant experience
• Excellent object-oriented design skills using C++
• Demonstrated experience in compiler implementation
• Strong experience in writing parallel programs, such as for GPUs, or multi-threaded or multi-process programs
• Strong written and communication skills
• Good understanding of parallel computations