About The Role In this role, you will contribute to the design, development, and optimization of high-performance GPU-accelerated algorithms for advanced numerical methods and computer vision applications. You will work closely with domain experts and engineering teams to translate complex mathematical models into efficient, scalable, and vector-friendly solutions, moving ideas from Python prototypes to production-grade C++ implementations. You will play an important role in improving application performance on modern NVIDIA GPU architectures, influencing technical decisions around CUDA kernel optimization, mixed-precision computing, task-based runtimes, and multi-GPU execution. Responsibilities Design, implement, and tune high-performance CUDA kernels Profile, benchmark, and iterate using Nsys, Nsight Collaborate with domain scientists to translate mathematical models into vector-friendly algorithms Prototype new approaches in Python and harden them in C++ Influence architecture decisions related to task-based runtimes, mixed-precision arithmetic, and multi-GPU scheduling Requirements Strong hands‑on experience with algorithmic design and optimization of numerical methods, with at least 12 months of commercial experience in this area Solid understanding of designing, developing, and profiling CUDA kernels for modern architectures: Ampere and above, CUDA core C++ libraries: CUB, Thrust, etc. Experience with C++ Background in Computer Vision Advanced knowledge in Microsoft Visual Studio SoftServe is an equal opportunity employer. Qualified applicants will receive consideration regardless of race, color, ancestry, ethnicity, national origin, religion, sex, sexual orientation, gender identity or expression, age, citizenship, disability, health condition, marital or family status, veteran status, or any other characteristic protected by applicable law. #J-18808-Ljbffr