AI SW Stack Deployment Architect

Responsibilities

• Architect integration of vLLM, PyTorch, and TensorFlow, JAX/XLA into Next Generation Accelerator stack
• Define framework → compiler → runtime APIs and contracts
• Own LLM execution behavior (batching, KV cache, streaming inference)
• Design and implement end-to-end deployment workflows (packaging, versioning, reproducibility)
• Drive performance optimization across model → framework → runtime
• Work cross-functionally with compiler, runtime, and low-level SW teams
• Support customer workloads, model onboarding, and debugging
• Impact
• Own customer-visible AI execution and deployment on Next Generation Accelerator , closing the gap between models and system performance , and enabling enterprise-grade AI solutions
• Required Qualifications
• 10+ years in AI/ML systems or software architecture
• Strong experience with PyTorch / Transformers / LLMs
• Hands-on experience with LLM deployment and scalable inference engine systems e.g. vLLM, Triton, SGLang etc.
• Experience building scalable AI platforms (cloud/edge)
• Expertise in system design, APIs, and cross-layer integration

Requirements

• Experience with vLLM or similar LLM serving systems
• Familiarity with XLA / MLIR / compiler frameworks
• Exposure to AI accelerators (GPU/NPU) and runtime systems
• Experience in distributed or multi-agent AI systems

Additional Information

Role Overview We are looking for a Software Architect (12+ years experience) to lead the application/framework layer and deployment stack for the Next Generation Accelerator AI platform. This role owns how models run on Next Generation Accelerator-from vLLM / PyTorch / TensotFlow/XLA to production deployment-ensuring correctness, performance, and scalability.

Your Match

How well this role fits your profile.

Company Intel

What employees say

Worked at Sandisk? Share your experience