Artificial Intelligence Engineer

VIVA Tech Talent; Belfast City District; On-site; GBP 90,000 - 130,000; Full time

Job summary

VIVA Tech Talent is hiring a senior AI engineer to lead the creation of an AI function from scratch. You will train models in Python, deploy agentic workflows to secure production on Azure, and build search and retrieval layers underneath them.

You will own the full modelling lifecycle, decide what to train on, and ensure inputs are secure and outputs verifiable, working closely with an engineering team and CTO on day one.

Qualifications

• Production-standard Python, with hands-on modelling in ML frameworks (PyTorch, TensorFlow or scikit-learn).
• Ownership of the full modelling lifecycle, from data sourcing/cleaning to encoding/vectorisation.
• Agentic workflows you’ve taken to production and supported, not prototypes on frontier models.
• Model security applied in production, including jailbreak prevention and prompt-injection defence.
• Production experience on a major cloud platform (Azure preferred; AWS Bedrock depth also acceptable).
• Vector databases used in production and retrieval systems that scale with data growth.
• Hands-on experience with MCP or an equivalent agent-tooling protocol.

Responsibilities

• Train models in Python and advance production-ready pipelines.
• Deploy agentic workflows to secure production on Azure, with scalable retrieval layers.
• Design what to train on and how to measure outputs for correctness.
• Collaborate with an engineering team and CTO to shape the platform’s direction.

Skills

Python

PyTorch

Scikit-learn

TensorFlow

Model security

Azure

Vector databases

Retrieval systems

Agent tooling protocol

End-to-end ML lifecycle

Tools

Azure

AWS Bedrock

Vector databases

MCP protocol

Job description

Most AI roles ask you to join a team. This one asks you to build it.

Our client is building an AI function from scratch, and the person they hire will shape it from day one, setting its direction and deciding what it takes on next.

It's a senior, hands-on role. You'll be training models in Python, taking agentic workflows into secure production on Azure, and building the search and retrieval layers underneath them. You'll also make the bigger calls. What to train on. How to secure every input that reaches a model. How to prove an output is correct, not just plausible.

You won't be doing it alone. There's an engineering team that has already built and shipped the platform, and a CTO close to the work.

What you'll need

• Production-standard Python, with hands-on modelling in PyTorch and at least one of sklearn or TensorFlow
• Ownership of the full modelling lifecycle, from sourcing and cleaning data through to encoding and vectorisation
• Agentic workflows you've taken into production and supported, not prototypes built on a frontier model
• Model security applied in production, including jailbreak prevention and prompt injection defence
• Production experience on a major cloud A.. platform. Azure is preferred, but equivalent depth on AWS Bedrock transfers
• Vector databases run in production, and retrieval systems that stayed accurate as the data grew
• Hands-on experience with MCP or an equivalent agent-tooling protocol