Senior CUDA C++ Software Engineer

Position: Senior CUDA|C++ Software Engineer

Location: Arlington, VA

Clearance: TS/SCI

Grey Matters Defense Solutions, LLC is a specialized firm in software development, data analytics algorithms, and advanced remote sensing technologies, tailored to meet the complex demands of defense and intelligence sectors. Staffed by a diverse range of professionals—from senior-level experts from organizations like the Defense Intelligence Agency (DIA), National Reconnaissance Office (NRO), Defense Advanced Research Projects Agency (DARPA), and the U.S. Armed Forces, to recent graduates and military recruits.

By integrating the skills of subject matter experts, analysts, software engineers, and data scientists,
Grey Matters Defense Solutions, LLC delivers unique artificial intelligence algorithms and applications, positioning itself as a leader in innovative solutions for defense and intelligence.

The Senior CUDA/C++ Software Engineer designs, develops, and optimizes high-performance GPU-accelerated software solutions supporting advanced signal processing, RF exploitation, and SIGINT mission applications. This position focuses on building scalable, production-quality software using C++, CUDA, and NVIDIA GPU technologies to process large volumes of data in real-time or near-real-time environments. The ideal candidate possesses deep expertise in modern C++ development, GPU architecture, CUDA optimization, and high-performance computing (HPC).

Experience with DSP, RF signal processing, and SIGINT applications is highly desirable.

8+ years of professional software engineering experience. 5+ years of experience developing high-performance applications in C++. 3+ years of hands-on CUDA development experience in production environments. Experience optimizing applications for throughput, latency, memory utilization, and scalability. Experience developing software on Linux platforms.

Design, develop, and maintain GPU-accelerated applications using CUDA and modern C++. Develop custom CUDA kernels optimized for performance, scalability, and maintainability. Implement efficient memory management strategies utilizing shared memory, pinned memory, unified memory, and asynchronous transfers. Optimize GPU utilization through concurrency, streams, kernel fusion, and multi-GPU processing techniques. Integrate GPU-accelerated capabilities into larger software systems and mission applications.

Analyze and optimize application performance across CPU and GPU architectures. Profile applications using NVIDIA Nsight and related performance analysis tools. Identify and resolve bottlenecks related to memory bandwidth, kernel execution, synchronization, and data movement. Develop benchmark frameworks and performance regression testing capabilities.

Develop clean, maintainable, and well-documented code using modern C++ standards. Participate in software architecture, design reviews, and code reviews. Build automated unit, integration, and performance testing frameworks. Support CI/CD pipelines and automated build environments. Collaborate with cross-functional engineering teams to deliver mission-critical software capabilities.

Bachelor’s degree or higher in Computer Science, Computer Engineering, Electrical Engineering, Applied Mathematics, Physics, or a related STEM discipline.

C++17/20 CUDA
Python
Linux Development Git
CUDA Kernel Development NVIDIA GPU Architecture CUDA Streams and Concurrency Memory Optimization
Multi-GPU Processing Nsight Profiling cuFFT
cuBLAS
Object-Oriented Design Design Patterns
Unit Testing CI/CD
Performance Benchmarking Software Architecture Digital Signal Processing RF Systems
IQ Data Processing SIGINT Applications
High-Performance Computing

Grey Matters Defense Solutions, LLC provides transformational leadership, building aware-winning teams and products. Grey Matters Defense Solutions’ most valuable assets are the more than 60 + employees. Our employees have a depth of analytical knowledge which provides them with a deep understanding of managing and delivering products within government…