Principal AI Security Software Engineer

Source Technology Limited; Greater London; On-site; GBP 189,000 - 208,000; Full time

Job summary

Source Technology Limited in London is seeking a Principal AI Software Engineer to lead the hands-on delivery of a portfolio of agentic AI security applications. You will translate a complex target architecture into production-quality software, supervising multiple engineers while remaining deeply technical.

This role requires deep experience with agentic AI, LLM orchestration, DevSecOps and secure SDLC practices, plus AWS and IaC.

Qualifications

• Experience leading engineering teams and delivering production-grade systems.
• Hands-on with agentic AI apps: LLM orchestration, retrieval, tool calling.
• Experience with AWS and IaC, and secure SDLC practices.

Responsibilities

• Lead the build of agentic AI security apps and define target architectures.
• Enforce secure SDLC and integrate with CI/CD platforms.
• Mentor engineers and raise engineering standards.
• Create reusable frameworks and platform capabilities.
• Deliver complex components across the agentic roadmap.

Skills

AI leadership

Agentic AI apps

Python/TypeScript

DevSecOps

CI/CD pipelines

AWS cloud

Security by default

Tools

Terraform

Docker

Kubernetes

GitHub Actions

Job description

Principal AI Software Engineer

£725-800 per day

London x1 day per week in St Pauls

6 month initial contract Inside IR35

The opportunity

This is a contract engagement supporting a major Information Security AI transformation, a large programme that is modernising and automating the security operating model with agentic AI.

We are looking for a hands‑on engineering leader to drive the build of a portfolio of agentic AI security applications, defanging the solution architecture and contributing to the most complex components of the solution and system design. You would work with a good deal of autonomy to deliver a defined set of outcomes, on complex and current problems in AI security, and collaborate with software and solution architects to define target architectures, and coordinate workloads of software engineers as well as being hands on.

What you would deliver

• Technical leadership of the build, turning the target architecture into production quality software across a portfolio of agentic AI applications, APIs and services.
• Engineering standards, DevSecOps practice and a secure development lifecycle, integrated with the organisation’s established central CI/CD platform, pipelines and golden paths.
• An AI-native software delivery operating model built on agentic harnesses and automation.
• Reusable frameworks, libraries and platform capabilities that speed up delivery and keep security built in by default.
• Delivery of the most complex components and integrations across our agentic roadmap.
• Technical assurance through code review, quality standards and mentoring, to raise the engineering bar.

Skills and experience

• Extensive practical software engineering experience, including leading engineering teams technically and delivering complex, production quality systems at scale.
• Deep, current experience building agentic AI applications, including LLM orchestration, retrieval augmented generation, tool and function calling, MCP, memory, evaluation and agent harnesses, in Python or TypeScript.
• Demonstrable knowledge of Langgraph agentic development frameworks.
• A track record of implementing or evolving engineering standards, DevSecOps practice and a secure development lifecycle, and integrating with a central CI/CD platform, pipelines and golden paths.
• Experience defining or evolving an engineering operating model and ways of working, ideally an AI native model built on agentic harnesses and automation, or software factory concepts.
• Experience delivering on cloud platforms, ideally AWS, with containers and Infrastructure as Code.
• A solid grounding in secure by default engineering, observability and production operability, with practical knowledge of AI security practices such as the OWASP Top 10 for LLM and Generative AI applications and OWASP guidance for agentic AI.

We believe in equal opportunity for all and actively encourage applications from diverse backgrounds, experiences, and perspectives.