Teams at Hybridizer

Own and develop HPC/GPU infrastructure for R&D and customer deployments. Design cross-platform CI/CD pipelines, manage on-premise and cloud Kubernetes clusters with GPU passthrough and MIG, configure self-hosted GitHub Actions runners, and maintain DockerHub registries. Assemble and tune physical GPU servers, manage PCIe topology and hardware constraints, and maintain compatibility across NVIDIA drivers, CUDA, and ROCm versions. Provide customer-facing technical assistance for GPU container deployments.

Develop compiler and high-performance computing solutions that translate .NET and Java bytecode into optimized CUDA and ROCm code for CPUs and GPUs. Analyze customer algorithms, design parallel implementations, extend compilation infrastructure, build reusable libraries, and optimize throughput and latency. Apply expertise in GPU programming, compiler architecture, mathematics, and software engineering to deliver performance matching or exceeding hand-tuned binaries.

R&D intern developing and validating compiler features for high-performance computing. Responsibilities include analyzing CPU/GPU performance, reading PTX and LLVM assembly, implementing hardware intrinsics and mathematical algorithms, fixing compiler bugs, creating benchmarks and demos, and writing technical documentation. The role requires strong C++ or C# skills, understanding of memory management and object-oriented programming, and interest in CUDA, OpenMP, GPU architecture, algorithms, and numerical methods.