AI Systems Engineer – Custom AI Chip Integration

About the role

Microsoft's Superintelligence team is building next‑generation AI silicon and needs a senior engineer to bridge the gap between custom hardware and modern AI inference frameworks. You will own the software integration layer that enables our proprietary AI chip to run large language model workloads with peak efficiency.

Key responsibilities

• Architect and develop backend integration so the custom AI chip becomes a first‑class citizen in the SGLang serving framework.
• Write custom C++/PyTorch extensions that map SGLang primitives (e.g., RadixAttention, FlashAttention, matrix multiplications) to the chip’s proprietary SDK.
• Profile and optimize end‑to‑end inference latency, throughput, and memory utilization on the hardware.
• Collaborate with hardware architecture and compiler teams to provide feedback on software stack and silicon design.
• Build robust testing and deployment pipelines for the integration.

Required profile

• Proven experience in systems engineering for AI hardware or high‑performance computing.
• Strong bias toward real‑world impact and end‑to‑end ownership.
• Ability to work cross‑functionally with hardware, compiler, and software teams.

Required skills

• C++ programming
• PyTorch extension development
• SGLang framework familiarity
• Custom AI chip SDK integration
• Performance profiling and optimization
• Knowledge of attention mechanisms such as FlashAttention and RadixAttention

What we offer

• Opportunity to shape the future of AI infrastructure at a leading technology company.
• Collaborative, inclusive culture with a growth mindset.
• Access to cutting‑edge AI hardware and research projects.