Software Security Researcher, GPU and AI Systems

Role overview

This role centers on performing original security research across the GPU and AI infrastructure stack and converting findings into runtime detections that are deployed into real environments. The scope spans low-level driver and runtime layers, container and virtualization boundaries, multi-tenant isolation models, and the integrity of machine learning model supply chains. The position is research-driven but emphasizes shipped impact rather than purely academic output.

Responsibilities

- Conduct vulnerability and weakness research across GPU drivers, AI runtimes, and adjacent infrastructure layers - Investigate container and virtualization boundaries for isolation, privilege, and escape issues - Analyze multi-tenant GPU and AI compute environments for cross-tenant and cross-workload risk - Research threats targeting AI/ML model supply chains, including model loading paths and artifact integrity - Author runtime detections that operationalize research findings - Partner with engineering teams to integrate detections into deployed systems

Requirements

- Demonstrated experience conducting security research on systems-level targets such as drivers, runtimes, kernels, or hypervisors - Familiarity with GPU compute architectures and modern AI/ML frameworks and their underlying mechanics - Understanding of container and virtualization internals and the security boundaries they create or fail to enforce - Strong systems programming skills for building reliable, performant detection logic - Ability to translate abstract research findings into concrete, observable runtime controls that scale in production

Nice to have

- Hands-on experience with multi-tenant GPU or AI infrastructure in cloud or bare-metal settings - Background in offensive techniques, reverse engineering, fuzzing, or program analysis - Prior publications, CVEs, bug bounty reports, or open-source contributions in security research