CUDA Engineering Expert

Position:
CUDA Engineering Expert

Type:
Hourly Contract

Compensation:
$80–$100/hour

Location:
Remote

About the Opportunity

This opportunity is for experienced
GPU Performance Engineers, CUDA Developers, and GPU Kernel Optimization Specialists
interested in contributing to advanced AI research and evaluation projects.

The role focuses on analyzing, optimizing, and evaluating GPU kernels across modern hardware architectures. You'll leverage your expertise in
CUDA, C++17, GPU profiling, and performance optimization
to improve computational efficiency and help advance next-generation AI systems.

This is a contract-based opportunity for professionals passionate about maximizing GPU performance and hardware utilization.

Responsibilities
- Analyze and optimize GPU kernels for
performance, efficiency, scalability, and hardware utilization
.
- Use profiler metrics such as
L2 cache hit rate, occupancy, memory throughput, warp efficiency, and related performance indicators
to guide optimization decisions.




- Identify bottlenecks in GPU kernel implementations and recommend performance improvements.
- Develop, review, and optimize
C++17, Python, and GPU programming code
.
- Apply expertise in
CUDA, HIP, shader programming, or related GPU programming frameworks
to improve kernel performance.
- Document optimization methodologies, profiling results, and engineering decisions with clear technical reasoning.
- Collaborate with engineering teams to evaluate and improve AI-related GPU workloads.

Required Qualifications
- Availability to work
at least 20 hours per week
.
- Solid proficiency in
C++ (through C++17)
.
- Working knowledge of
Python
and
Git
.
- Professional experience with at least one GPU programming framework, including
CUDA, HIP, Slang, HLSL, GLSL, or similar technologies
.
- 1+ year of professional or graduate-level research experience
working with GPU programming or optimization.
- Strong understanding of GPU architecture and performan

📌 Hardware Engineer (India)
🏢 Synthires
📍 India