Data Center Security Architect

Lucid Computing•San Francisco, CA

About The Position

Lucid Computing, in partnership with the Research Institutes of Sweden (RISE), is building and operating an adversarial testbed to develop and red-team AI compute verification solutions. The role develops the physical and operational security layer underlying verification designs, helping protect in-path components and their guarantees against bypass, substitution, tampering, and the introduction of undeclared hardware. Security measures may include cameras, custom server enclosures, tamper-evident mechanisms, random spot inspections, countermeasures against undeclared wireless components, controlled installation and maintenance procedures, and chain-of-custody processes for security-relevant devices. This is not a conventional data center security role focused on badge programs or static compliance checklists. The successful candidate will design security protocols for a threat model involving a well-resourced, covert adversary that may include the facility operator. Designs will be tested by an independent red team, with results—including negative findings—published under a pre-agreed protocol with RISE. The candidate will be the physical and operational security specialist on a multidisciplinary team of hardware and network engineers, cryptographers, ML researchers, and governance researchers. They will contribute original security designs, implement and operate protocols in experiments on a real-world GPU cluster, and oversee secure installation of in-path verification devices alongside live workloads. They will also produce installation-assurance documentation, inspection methodologies, and physical threat models for publication in an open reference architecture. The work aims to inform how governments and AI labs approach verification and support coordination around AI security policies and standards. The role can be based in San Francisco or London, or fully remote. The posting states that compensation includes a base salary within a range listed above and equity, but does not provide the salary figures in the supplied description.

Requirements

Hands-on experience securing operational AI/GPU data centers, including designing, implementing, or running physical and operational security for facilities hosting large-scale accelerated compute.
Deep expertise in physical security controls and operations, including access control and zoning, surveillance and camera coverage, intrusion detection, tamper-evident seals and enclosures, secure racks or cages, visitor and vendor escort procedures, and guard or inspection operations.
Experience threat-modeling sophisticated adversaries, including insider threats and well-resourced external actors.
Working fluency in data center infrastructure, including power, cooling, structured cabling, network topology, and out-of-band management, sufficient to collaborate with network and hardware engineers and identify interactions between physical and logical security assumptions.
Experience conducting, supporting, or remediating findings from physical penetration tests, red-team exercises, or hostile facility audits.
Strong security writing skills, including the ability to produce clear threat models, SOPs, inspection protocols, and assurance documentation that third parties can execute and external experts can critique.
Demonstrated creativity and first-principles thinking, with evidence of having designed or being capable of designing novel security mechanisms.
Comfort with public scrutiny of security work and publication of results, including negative findings.

Nice To Haves

Experience securing hardware lifecycle and chain of custody, including secure receipt, provenance verification, installation, maintenance, and decommissioning of security-relevant hardware, with procedures designed to remain trustworthy when staff or vendors may not be trusted.
Experience in cleared, regulated, or high-assurance facilities, such as SCIF construction or accreditation (ICD 705 or national equivalents), TEMPEST or emanation security, defense or intelligence-community data centers, or financial-exchange-grade facilities.
Experience with side-channel and covert-channel defenses, including RF, acoustic, or power-line emanation assessment, shielding, filtering, jamming, or detection of undeclared wireless devices.
Hardware and firmware security knowledge, such as JTAG or SPI-flash attacks and defenses, BMC/DC-SCM security, firmware attestation, hardware roots of trust, or supply-chain attack analysis.
Background in arms-control or safeguards verification, such as IAEA-style inspection regimes, tamper-indicating devices, containment and surveillance, information barriers, or chain of custody in bilateral low-trust settings.
Familiarity with AI verification and compute governance literature, including workload classification, network taps, recomputation, and flexible hardware-enabled guarantees, or ability to learn the area quickly.
Experience with relevant standards and frameworks, including EN 50600, TIA-942, ISO 27001 physical and environmental controls, or SOC 2, and judgment about their limitations for this threat model.
Understanding of confidential computing and attestation, including TEEs and measured boot, sufficient to reason about how physical security composes with hardware-rooted evidence.
Experience designing insider-threat and personnel-security programs, including two-person integrity, background or vetting processes, and diverse-team inspection protocols.
Publication or public-speaking experience, including the ability to represent work to labs, governments, and the research community.
Experience collaborating in research or R&D environments and comfort with ambiguity, iteration, and public critique.
Eligibility for, or history of holding, a national security clearance.
Enthusiasm for building security in the open and publishing results, including negative findings.

Responsibilities

Develop the physical and operational security layer of AI compute verification designs with a multidisciplinary team.
Design original security mechanisms and protocols for a threat model involving sophisticated covert adversaries, including potential facility-operator adversaries.
Assess how verification devices and other security-relevant components can be installed, protected, inspected, and maintained without relying on a trusted installer.
Develop approaches to detecting hidden or undeclared hardware, including unauthorized components inside GPU servers and undeclared wireless devices.
Design practical inspection regimes for declared local storage and other security-relevant equipment.
Implement and operate security protocols in experiments on a real-world GPU cluster.
Oversee secure installation of in-path verification devices alongside live workloads.
Develop installation-assurance documentation, inspection methodologies, and physical threat models for publication in an open reference architecture.
Engage with independent red-team testing and public scrutiny of designs and findings, including negative results.
Collaborate with hardware engineers, network engineers, cryptographers, ML researchers, and governance researchers.