Staff Engineer, Inference Optimizations

📋 Description Performance architecture: Benchmark and optimize inference engine and GPU kernels. Deep-dive optimization: Address bottlenecks in attention, memory, and parallelism. Technological innovation: Apply cutting-edge optimization techniques for Gen AI workloads. Hardware & ecosystem mastery: Advise on NVIDIA/AMD GPUs, CUDA, ROCm, and stacks. Precision optimization: Develop FP8/INT8/BF16/FP4 quantization to boost throughput. Technical mentorship: Lead code/design reviews and raise technical bar. 🎯 Requirements Technical Depth: 5+ years in HPC/AI infra; solve compute and memory bottlenecks. Gen AI Literacy: Deep familiarity with Gen AI landscape and model families. Optimization Expert: Hands-on with attention-layer optimizations and distributed GPU parallelization. Hardware Fluency: Deep understanding of NVIDIA/AMD GPUs and CUDA/ROCm stacks. Open Source Mastery: Experience building with and contributing to open-source. Systems Design: Low-level GPU programming, memory access patterns and parallel execution. 🎁 Benefits Innovate with purpose; think big, bold, and act like an owner. Prioritize career development with conferences, training, and LinkedIn Learning. Care for well-being with competitive benefits and flexible time off. Salary, bonus opportunities, equity, and Employee Stock Purchase Program. Cobalt Streamworks is an equal-opportunity employer; no discrimination.