Job Vacancies

Bachelor's degree (Graduate) in Computer Science, Software Engineering, Artificial Intelligence, Data Science, or equivalent.

3-5 years total professional experience in software, automation or platform engineering.

Within that, at least 1-2 years hands-on experience building AI agents or LLM- based automation solutions that reached real users — proof-of-concept and demo work alone is not sufficient.

Evidence of at least one AI or automation solution taken into day-to-day operational use, with the candidate able to describe how it was tested, what failed in practice, and how it was corrected.

Experience deploying solutions within an enterprise environment subject to change control and security review.

A. Agent Design & Development  Design, develop and deploy AI agents that automate defined business, project, compliance and administrative workflows.  Select the appropriate approach for each problem — a tool-using agent, a multi- step agenti

B. Enterprise Integration ï‚· Build and maintain integrations between AI agents and MVL systems, including the ERP and project management platforms, Microsoft 365, SharePoint Online, Teams and Dataverse. ï‚· Implement integrations to production stand

C. Retrieval & Knowledge Engineering ï‚· Design and operate retrieval-augmented generation (RAG) pipelines over MVL's document estate, including project documentation, contracts, policies and correspondence.ï‚· Own the ingestion pipeline: format hand

D. Evaluation, Testing & Quality ï‚· Build and maintain test sets and evaluation checks for every production agent, covering both retrieval quality and end-to-end task success. ï‚· Define the accuracy threshold an agent must meet before it is release

E. Deployment, Operations & Monitoring ï‚· Deploy agents onto MVL's approved Azure and Microsoft 365 platforms using repeatable, source-controlled pipelines rather than manual configuration. ï‚· Instrument agent runs so that tool calls, retrieved con

F. Security, Governance & Documentation ï‚· Apply AI-specific security controls, including defence against direct and indirect prompt injection, restriction of the tools an agent may invoke, and validation of outputs before any action is taken. ï‚· T