Technical Lead – AI & Software Engineering

Job Title:

Technical Lead – AI & Software Engineering

Salary: £53,100 - £70,950

Location:

Cambridge, UK - Hybrid (at least 2-3 days per week in the office)

Contract:

Permanent

Hours:

Full-time, 35 hours per week

We are recruiting a Tech Lead for each of two projects. Each Tech Lead will work within a POD led by the Technical Product Manager for their project, reporting to the Solutions Lead.

We are Cambridge University Press & Assessment, a world-leading academic publisher and assessment organisation and a proud part of the University of Cambridge.

We're seeking a hands-on Tech Lead with strong leadership and collaboration skills to lead delivery of AI-enabled solutions that transform how assessments are created and marked. Your solutions will depend on privately hosted language models, and you'll ensure they are secure, reliable, auditable and grounded in approved assessment content, working within the enterprise direction set by the Solution Lead.

In this role, you will not only lead and mentor a small cross‑functional squad, but also remain deeply involved in the technical delivery—applying your expertise directly to architect, code, and implement solutions. You will set technical standards, write and review code, and solve complex challenges alongside your team, taking a first release from idea to live use by assessment experts and examiners.

Throughout, you'll work closely with those experts, who remain in control of every decision that affects a candidate. We are looking for someone who leads by example, demonstrates recent hands‑on experience of building products with language models, and is passionate about both technical excellence and team development.

• Content Creation: AI‑assisted solutions that help to transform the authoring and quality assurance of assessment content, including question papers and mark schemes.
• Marking and Core Operations: AI‑assisted tools that support examiners in reading handwritten scripts and marking consistently, with examiners making every decision.

• Deliver AI‑Enabled Solutions:
  Build end‑to‑end products that use securely hosted language models, from data and retrieval through to user interfaces and integration with existing assessment systems, while communicating design rationale to diverse stakeholders for buy‑in and feedback.
• Ground and Validate Outputs:
  Use retrieval over approved assessment content, structured outputs and rule‑based validation so that results are accurate, traceable and consistent with syllabus requirements, working with assessment experts owners to agree what “correct” looks like.
• Measure Quality:
  Build evaluation sets and quality gates alongside assessment experts, using them to select and compare models and to show each release is fit for use, fostering a team culture of shared accountability and continuous improvement.
• Keep Experts in Control:
  Design review workflows with examiners and assessment experts so they make every final decision, with every AI suggestion and human decision recorded for audit.
• Protect Sensitive Content:
  Champion secure practices aligned with industry standards, ensuring models, data, prompts and logs stay within approved secure environments, and influencing stakeholders to prioritise security in decision‑making.
• Shape Cloud‑Native Architecture:
  Design scalable, resilient architectures on AWS using approaches like microservices, serverless, containerisation and GPU‑based model hosting, using Infrastructure as Code for cost‑effective, repeatable deployments.
• Drive Observability:
  Implement robust observability covering system performance, reliability and the quality of AI outputs, so issues are found and resolved proactively.
• Integrate with our…