Applied AI Engineer

This is a new anda senior role forCloud

Guard'sapplied AI function in Ireland.

Location: Hybrid — with 1 day per week in Belfast Office

Employment Type: Full-Time, Permanent

Salary: Competitive — dependent on experience

This is a new anda senior role. We are standing upCloud

Guard'sapplied AI function and the person we hire builds it and sets its direction.

Day to day you will be hands-on in Python and Azure — training models, taking agentic workflows from prototype into secure production deployment, and building the search and retrieval layers underneath them.

You will also be trusted with the wider decisions that come with the role: what we train on, how we secure every input surface that reaches a model, and how we satisfy ourselves that an output is correct rather than plausible. We are asking for experience because you will be making those calls as you grow the role. You will have people to think with: an engineering team that has built and shipped this platform, analysts who run our SOC every day, and a CTO close to the work.

• Build and train ML models in Python using
  
  PyTorch,sklearnand Tensor Flow, and own the data work behind them: sourcing, cleaning, encoding andvectorisation.
• Take agentic workflows from prototype into secure, production-grade deployment, and support them once they arelive.
• Design andmaintainsearch and retrieval systems as our data needs grow.
• Build on Azure as the platform’s primary AI tooling.

Security & Model Protection

• Apply model security across every input surface: jailbreak prevention and prompt injectiondefencewherever untrusted input reaches a model.
• Extend the platform’s model guardrails as you add new capability.
• Flag risk early — "here's the gotcha" or "consider doing it this way" — rather than waiting to be asked.

Leadership & Strategic Input

• Own the technical direction ofCloud
  
  Guard'sapplied AI function, including which problems we take on next.
• Advise onour data strategy and on the architecture of the secure interfaces that sit in front of our models.
• Validate and improve existing AI capability so we can stand behind what it produces.
• Act as a second, deeper line of technical judgement alongside the engineers already reviewing this work.

Cloud Guard is a UK-based Microsoft Security specialist MSSP – built differently from the ground up. Operating dual Security Operations Centres in Manchester and Belfast, we combine deep Microsoft security expertise with human-led operations to deliver security that is measurable, scalable, and continuously evolving. We exist to close the gap between the security organisations think they have and the security they actually need.

‘Security. Done Different.’ is not a tagline, it’s how we operate every day.

Experience & Background

• Python at production standard, with hands‑on modelling experience inPyTorchand at least one ofsklearnor Tensor Flow.
• You have owned the full modelling lifecycle yourself: sourcing, cleaning, encoding andvectorisation, including how raw text becomes something a model can use and the decisions that get baked in at that first layer.
• Agentic workflows you have taken into production and then supported, rather than prototypes built against a frontier model.
• Model security you have applied in production: jailbreak prevention and prompt injectiondefenceacross every point where untrusted input reaches a model.
• Production experience building agents,workflows and model deployments on a major cloud AI platform. Ours is Azure, and equivalent depth on AWS Bedrock or similar transfers.
• Vector databases you have run in production, and search and retrieval systems that stayedaccurateas the data behind them grew.
• Hands‑on with MCP or an equivalent agent‑tooling…