N
Senior Software Architect, AI and GPU Networking
PresencialTiempo completoSoftware ArchitectureComputer HardwareEnterprise Software
Lead architecture and development for AI networking technologies that accelerate workloads in modern data centers. The role combines system research, hands-on software development, hardware co-design, and proof-of-concept delivery.
Responsabilidades
- Enhance GPU networking technologies for AI workloads, including runtime and communication solutions.
- Co-design GPU, DPU, and interconnect hardware features that accelerate data movement and enable inference and model-serving capabilities.
- Research and evaluate technologies, innovations, and partnerships against technical roadmaps and business value.
- Lead architecture and design for runtime systems, communication libraries, and AI-specific technologies.
- Lead proof-of-concept development to assess and advance new technologies.
- Explore transport functions, AI system communication, distributed AI, deep learning, HPC, software-defined networking, virtualization, and storage.
Requisitos
- M.Sc. or Ph.D. in Computer Science, Electrical Engineering, or Computer Engineering, or equivalent experience.
- At least 5 years of industry experience in system architecture, AI systems architecture, AI scaling, AI framework parallelism, or deep learning training workloads.
- Experience with algorithm design, system programming, computer architecture, and operating systems.
- Experience with virtualization, networking, and storage.
- Strong understanding of performance profiling, optimization, and hardware feature definition and use.
- Strong programming and software development skills.
- Ability to work and communicate effectively across multinational, multi-time-zone teams.
Se valora
- Research track record.
- Experience with CPU, GPU, memory, storage, or networking architecture.
- Knowledge of deep learning frameworks and AI communication libraries such as NCCL, UCX, or MPI.
- Understanding of inference and training workload optimization, including prefill/decode and parallelism techniques.
- Strong communication skills.