CUDA Developer – AI/LLM GPU Optimization

About the role

Gramian Consultancy is seeking an experienced CUDA Developer to join a 3‑month contract focused on advancing AI and large language model capabilities. The role is fully remote and requires close collaboration with AI researchers and engineers to solve complex GPU programming challenges.

Key responsibilities

• Solve advanced CUDA and GPU programming problems, optimizing parallel workloads for performance.
• Review, evaluate, and improve AI‑generated CUDA, C++, and Python code.
• Optimize GPU kernels for throughput, latency, memory efficiency, and resource utilization.
• Utilize CUDA libraries and frameworks such as Thrust, cuBLAS, and cuDNN.
• Debug and resolve issues related to kernel execution, synchronization, and memory management.
• Develop high‑quality technical prompts, solutions, explanations, and evaluations for AI model training.
• Collaborate with AI researchers, engineers, and evaluation teams.

Required profile

• 5+ years of professional software development experience with a strong focus on CUDA.
• Proficiency in C/C++ and hands‑on experience with Python and scientific computing ecosystems.
• Experience with PyTorch and NumPy.
• Deep understanding of GPU programming, parallel computing, and performance optimization.
• Familiarity with CUDA 12.3 or newer and related libraries (Thrust, cuBLAS, cuDNN).

Required skills

• CUDA (12.3+)
• C/C++
• Python
• PyTorch
• NumPy
• Thrust
• cuBLAS
• cuDNN