Staff Software Engineer - AI Accelerator Card Security Engineer - Cork, Ireland
Qualcomm
- Location
- Cork, Ireland
- Work model
- On-Site
- Level
- Staff
- H-1B history
- 22 approvals (FY2023)
- Posted
- Sep 17, 2026
Skills
About this role
Company: QT Technologies Ireland Limited Job Area: Engineering Group, Engineering Group > Software Engineering General Summary: Qualcomm Technology Inc.’s Dragonfly AI accelerator cards provide the backbone of an enhanced AI experience by shifting the AI workloads away from generic processors and onto specialized NPU cores that provide faster execution times, shorter latencies, and lower power consumption. These accelerators bear the responsibility of processing valuable user and model data – meaning that security is a critical product requirement that must be achieved to guarantee the products’ success in the market. We are seeking a talented security engineer to join the Dragonfly security team to ensure that the components of the Dragonfly AI accelerator card meet the security requirements and are securely configured and provisioned. You will determine the appropriate security requirements for the components on an AI accelerator card; evaluate third-party components; determine appropriate strategies for provisioning security information to the devices; and work with our partners to ensure that these strategies are implemented.
Responsibilities
In this role, you will be responsible for ensuring that the components and software on the Dragonfly AI accelerator cards meet security requirements, and are securely configured and provisioned: Define and document the security requirements for components on an AI Accelerator card. Define and document the mechanisms used for secure communication between the Board Management Controller (BMC) and the platform-management services, including relevant Redfish, SPDM-based, and PLDM-based interactions. Define and document processes to establish the device identity, collect measurements, generate evidence, and verify that evidence that allow a Platform Root of Trust (PRoT) to attest the state of the AI accelerator card to a management service through the BMC. Represent the Dragonfly security team in relevant customer, vendor, and standards forums, especially forums related to the use, development, and interoperability of attestation protocols. Evaluate components for their ability to meet the security requirements and work with vendors to improve components that do not meet the relevant requirements. Design and document services for device-personalization and provisioning of keys, certificates, identities, and configuration data in factory environments. Design and document test plans and production-line validation systems to verify that identities, keys, certificates, and secure configuration data have been provisioned correctly to devices in the factory environment. The role combines security architecture and requirements development with hands-on prototyping, implementation review, automation, and validation. Critical Skills: The successful candidate will have the following skills: A Bachelor's degree in Engineering, Information Systems, Computer Science, or related field, and 7+ years of experience in product security or a related field. Work experience gained by successfully completing a Master’s or Doctoral degree will be taken into consideration. An understanding of the security relevance of the physical components that exist on an AI accelerator card (such as the BMC, the PRoT, flash devices, and PCIe switches, including the use of ACS). An understanding of the principles of key management and certificate management. An understanding of the principles of the security lifecycle of devices, root-of-trust security, boot-chain security, and firmware authentication. Experience in performing security risk assessments on systems of discrete components. Experience in drafting security specifications and reviewing systems to determine if they meet those specifications. Proficiency in the security review of firmware and software that executes on the physical components on an AI accelerator card written in the C programming language. Proficiency in the use of the Python programming