Member of Technical Staff - Security Engineering

Description

The company was founded in 2022 to address how the time needed to produce useful outputs limits the impact of the most capable AI systems. It builds chips and systems that generate output at thousands of tokens per second while handling very large models at very long contexts.

Frontier workloads already deliver real value. Technical and economic limits on inference speed are what hold back progress in drug discovery, software engineering, materials discovery and other fields that depend on deep reasoning.

The company is young and works across hardware and software. It recently raised $220M from investors including Founders Fund and Accel.

You would work on the software layer of a full inference system, which is still taking shape. Several design decisions remain open, and you would help settle them.

Titles matter less than experience. Senior, staff, principal or lead all fit, provided you bring deep security judgement, system-level experience and good teamwork.

You would join for first bring-up: the first secure boot that verifies, the first device that attests itself, and the first clean end-to-end run.

Responsibilities

• Build the control chain that powers on, monitors, updates and protects devices and racks across bare metal, RTOS and embedded Linux.
• Build the communication mechanisms that connect devices to each other and to the host.
• Carry security through the whole chain, from the hardware root of trust up to the rack management plane, covering boot, signed updates, device identity, authenticated channels and runtime protection of the host Linux kernel and on-die firmware.
• Turn threat models into working mechanisms in bare metal, RTOS and Linux, then design, prototype and validate them.
• Shape the security of the control chain across device, board, chassis and rack controllers, including threat models, trust boundaries and a secure-by-design approach on every platform.
• Design and build secure boot and the root of trust on each controller: verified or measured boot, signing and verification, anti-rollback, and locking down debug and privileged access in production.
• Authenticate and protect the interconnects between controllers and between devices and the host, adding integrity protection where needed.
• Design the cryptographic mechanisms: key hierarchies, provisioning and storage, device identity, and remote attestation so the fleet can prove what it runs.
• Make firmware updates safe at every hop: signed images, rollback protection, staged rollout and clean recovery.
• Harden the host Linux kernel and on-die firmware once the system is running.
• Work with the architecture and hardware teams from the software and system view, and with the wider software team on threat-model reviews, vulnerability triage and driving fixes to closure.

Requirements

• Have an adversarial mindset and sharp threat-modeling instincts, covering trust boundaries, assets, attacks, and an honest view of residual risk.
• Have hands-on experience with the core security primitives in software and silicon, including secure boot, root of trust, cryptography and key management, and know how they are implemented and where they can fail.
• Understand how the root of trust, boot chain, key hierarchy and secure interconnects work underneath, and know which protections belong in silicon and which belong in software.
• Have designed secure protocols and channels between components.
• Write strong C and Rust, with a habit of memory safety and defensive coding.
• Have built security mechanisms yourself, such as a secure-boot chain, key hierarchy or attestation flow you shaped, not only used them.
• Be comfortable with ambiguity and bias toward action, since you will often secure systems before the hardware exists, working from a specification, then a simulator, then early boards.
• Not required, but especially welcome:
• Attestation at boot and at runtime, and management-plane protocols such as DICE, IETF RATS, SPDM, MCTP and PLDM, plus confidential-computing work such as TDISP.
• Hardware building blocks and firmware or Linux hardening, such as HSM, TPM, secure element, Caliptra and secure debug, open firmware or boot projects, and image minimisation.
• Assurance and resistance work: supply-chain security with SBOM and SLSA, fuzzing and testing, standards such as NIST 800-series, Common Criteria, OCP including S.A.F.E., and TCG or DMTF, plus physical and side-channel awareness.
• AI or LLM-assisted security tooling, and RISC-V, GPU or ML-accelerator experience.

Conditions

Applicants from every background are welcome, including those whose path into the field has not been linear.

The work involves technologies under UK, US and other international export controls, so some roles may need extra eligibility checks, which will be explained during hiring.

Benefits

• Competitive salary matched to your experience and the specialist nature of the role.
• Meaningful equity, so everyone shares in the value created.
• Health cover including medical, dental and vision, plus a contributory pension.
• Cover for life and critical illness.
• 25 days of annual leave on top of bank holidays.