Fal AI Senior Infrastructure Security Engineer at Fal AI securing the compute platform below the application layer, including Linux kernel, hypervisors, GPU drivers, and firmware. Focus on hardening the stack and improving multi-tenant isolation.
Responsibilities
We’re hiring a Senior Infrastructure Security Engineer to secure the foundation of fal.ai’s compute platform. You’ll own security below the application layer: the Linux kernel, the hypervisors and sandboxes that isolate workloads, GPU drivers and firmware, and the hosts that run them. You’ll harden the stack, strengthen isolation against untrusted code, and help make GPU compute safe at scale. Proficiency in building with and using AI is central to this role.
Secure Multi-Tenant Isolation
Build and maintain the threat model for running code on shared GPU hosts, from containers and microVM down to PCIe, IOMMU, and GPU passthrough
Design and verify GPU memory and device-state scrubbing between tenants
Help roll out hardware-backed isolation where it fits, including confidential computing and attestation
Harden the Stack
Harden kernel builds and host images: config minimization, LSMs, seccomp, namespace and cgroup policy, secure and measured boot
Reduce attack surface in GPU drivers, CUDA runtimes, NVLink/NVSwitch fabrics, BMC, and firmware
Assess and Research Vulnerabilities
Assess and triage security vulnerabilities in the kernel, hypervisors, GPU drivers, and firmware: judge exploitability and real impact on environment, set severity, and drive fixes or mitigations
Track vendor bulletins and upstream disclosures, and run kernel and driver patching that keeps pace without disrupting the fleet
Conduct vulnerability research on the components we depend on
Qualification
AI fluencyFluent in kernel CHas foundNVIDIA GPU stack experienceConfidential computing more broadlyPublished CVEsPrior work at a GPU cloud
Required
AI fluency: you use LLMs, agents, and AI coding tools daily to review code, build tooling, and accelerate vulnerability research
5+ years in systems or security engineering, with deep hands-on work in the Linux kernel, hypervisors, or device drivers
Fluent in kernel C, and comfortable in Rust, Go, or Python for tooling
Strong grasp of virtualization internals: KVM, VFIO, IOMMU/DMA protection, and the tradeoffs between containers, sandboxes, and microVMs
Has shipped hardening that runs in production at fleet scale
Has found, triaged, or fixed real vulnerabilities in low-level code and can explain the exploit path
Communicates risk clearly to different audiences, from kernel engineers to executives
NVIDIA GPU stack experience: open kernel modules, GSP firmware, MIG, vGPU, or GPU confidential computing
Confidential computing more broadly: AMD SEV-SNP, Intel TDX, attestation
Published CVEs, talks, or write-ups on kernel or hypervisor security
Prior work at a GPU cloud, hyperscaler, or chip vendor
Preferred
AI fluency: use LLMs, agents, and AI coding tools daily
