Senior Software Engineer / Architect – AI Inference Storage

Position Overview

We are looking for an experienced senior developer to design and build high-performance storage & networking systems optimized for AI inference workloads, particularly large language models (LLMs). This role involves developing scalable, GPU-accelerated solutions integrated with storage and network infrastructure that integrates tightly with modern AI inference frameworks.

Key Responsibilities

• Build RDMA data paths (RoCE/IB/iWARP) and integrate with the RDMA software stack.
• Implement GPU direct storage pipelines (NVIDIA GPUDirect, NIXL, future GPU-access technologies).
• Design and operate core components of a stateful distributed system (consensus, recovery, failover).
• Write Rust for high-performance components and Python for automation / AI integration.
• Advanced method in Linux system programming (async I/O, memory management, kernel interfaces).
• Prototype → production: own the path from design to real deployments.
• Mentor engineers and set strong coding and design standards.

Required Qualifications

• Deep experience with RDMA protocols and software stack internals.
• Proven work with GPU direct storage / GPUDirect / NIXL or similar direct GPU I/O.
• Strong Rust and Python development skills.
• Track record in distributed systems (stateful, fault tolerant, horizontally scalable).
• Advanced Linux systems programming knowledge.

Way to stand out

• Kubernetes for deploying and scaling stateful storage.
• Linux kernel programming (drivers, block layer, networking stack).
• Familiarity with AI frameworks (PyTorch, vLLM, TensorRT, Triton).
• Performance profiling, benchmarking at scale.
• Contributions to open-source projects.