Junior Solutions Architect (AI Assurance Engineer)

Junior Solutions Architect (Assurance Engineer, AI Safety) 

Please note that this role is subject to UK Security Check (SC) clearance. Applicants must be eligible to undergo and obtain SC clearance and have the right to work in the UK. Because vetting depends on individual circumstances, including background and residency history, we encourage applicants to review their likely eligibility before applying. 

Skills & Experience 

* Job role: Solutions Architect / Assurance Engineer 

* Experience Levels: Entry, Junior, Mid 

* Core Skills Considered: Python, Statistics & Mathematics, Git 

* Other Skills Considered: Docker, Kubernetes, AWS, LLMs, AI Risk Frameworks. 

About Advai 

Advai is building the AI Assurance infrastructure that sits between AI supply and AI demand - we aim to become the system through which organisations decide which AI to trust, deploy, and scale.  

Advai provides independent, automated assurance to choose, test, and monitor AI systems, accelerating AI adoption by generating the evidence required to deploy with confidence.  

Job Description 

We are looking for Junior Solutions Architects, aka ‘Assurance Engineers’, a new set of roles sitting at the forefront of understanding how AI can be tested, developed and deployed safely and securely.  

You will combine technical understanding of AI systems and human interpretation of their actions, helping build and execute evaluations for complex AI systems. In this role, you will work closely with our Senior Assurance Engineers to turn theoretical AI risks into practical, automated tests.  

You will contribute to translating client requirements and complex AI system architectures into structured, mathematically sound testing criteria. Once a testing plan is defined, you will be involved in the technical implementation and evaluation of client systems. This means writing the Python code to evaluate models, packaging those tests into Docker containers, and applying statistical rigour to ensure our safety evaluations are accurate. Interpreting client needs is a major skill in this field; you will actively participate in this process, learning how to map real-world business risks to frameworks like NIST and MITRE ATLAS, whilst gaining hands-on experience deploying your code into cloud environments. 

Responsibilities 

* Understand Risks: evaluating AI use cases and identifying potential risks 

* Test Design: Defining test cases and requirements for statistically sound, structured testing of AI systems. 

* Test Implementation: Write clean, well-documented Python code to execute performance, safety and security tests on AI models. 

* Evaluation & Analysis: Run evaluations on client AI systems, analyse the mathematical and statistical outputs, and help draft technical reports based on the findings. 

* Collaborative Research: Support senior engineers in threat modelling by reviewing academic literature on AI failure modes and red teaming. 

* Continuous Learning: Develop a deep understanding of AI risk frameworks (NIST, MITRE ATLAS, OWASP) and learn how to deploy testing infrastructure within AWS and Kubernetes (EKS). 

Minimum Qualifications 

* A degree in a quantitative subject (such as Mathematics, Computer Science, Physics, or similar), or equivalent practical experience. 

* Solid programming skills in Python, with the ability to write reproducible, structured code. 

* A strong foundation in statistics and mathematics, enabling you to understand how AI analysis is conducted and where its limitations lie. 

* A genuine, demonstrable interest in AI safety, robustness, and AI red teaming. 

* The ability to receive feedback and a proactive approach to solving novel problems. 

* Strong communication skills and a willingness to work collaboratively with a range of customers, understanding their needs. 

Optional Qualifications (Nice-to-Haves) 

* Basic experience using Docker or working with containerised applications. 

* Any exposure to cloud platforms (AWS, Azure) or Kubernetes. 

* Familiarity with machine learning libraries (e.g., PyTorch, TensorFlow, or scikit-learn). 

* Experience using Linux operating systems and command-line interfaces.