Purpose of the Role
The Applied AI Engineer supports POSH's AI programme by contributing to the delivery of prioritised use cases — assisting in the development of AI agents and solutions built on the Group's enterprise AI platform. This person will develop hands-on experience across the full AI delivery lifecycle: from translating business requirements into working prototypes, to deployment and user adoption support.
This is not a purely technical role. Equal weight is placed on business engagement, user adoption, and solution delivery as on technical execution. Prior domain knowledge is not a requirement for this role. This person is expected to spend time with the various functional and operational teams in POSH — understanding how the business works, what problems matter, and what good looks like in practice — and to bring that understanding back into everything they build.
Key Responsibilities
Use Case Delivery & Business Engagement
- Work with POSH business functions to understand operational workflows, identify AI opportunities, and validate that proposed solutions address the right problems.
- Translate business requirements into technical specifications and working prototypes.
- Facilitate structured user acceptance testing sessions with operational teams. Presenting solutions, capturing feedback, and documenting gaps for iteration
- Manage expectations proactively on timelines, data dependencies, and scope.
- Develop HITL review interfaces and feedback mechanisms in close collaboration with domain experts.
- Support the handover of deployed solutions to operational owners with documentation, user guides, and structured walkthroughs.
Data & System Integration
- Assess data quality, validate schemas, and surface integration issues with the Business Excellence Team and data owners.
- Integrate AI solutions with POSH's / Group's enterprise systems using the integration standards and infrastructure.
- Flag aspirational data assumptions and system-dependent dependencies early for programme-level decision
AI Agent Development
- Configure and deploy AI agents on the enterprise platform — applying agentic design patterns including multi-agent architectures, tool-use, and autonomous workflow execution.
- Build and maintain RAG pipelines for POSH-specific knowledge bases: ingestion, chunking, embedding, vector indexing, and retrieval optimisation.
- Implement LLM integration — prompt engineering, structured output design, and iterative refinement for operational and regulatory quality standards
- Build document processing components: OCR pipeline integration, multimodal content extraction and pre-processing for AI consumption.
- Document and escalate platform gaps or POSH-specific requirements for COE review.
Testing, Quality Assurance & Continuous Improvement
- Assist in developing and maintaining AI evaluation frameworks for POSH use cases: accuracy benchmarking, hallucination detection, retrieval quality scoring, and regression testing across model updates.
- Conduct structured testing of agent behaviour before any production deployment, including edge cases, adversarial inputs, and failure mode analysis
- Monitor deployed solutions; escalate anomalies promptly and support resolution with clear documentation.
- Collaborate with the Group Data & Automation COE on platform standards and technical governance.
- Document all agents, configurations, data schemas, and integration points to enterprise standards.
Experience
- 1-2 years of software engineering or data/AI engineering experience – including exposure to LLM-based development in any context. Internships, research attachments, and early-career roles are all considered.
- Hands-on experience with at least one AI or ML project from concept to working output. Production deployment is advantageous but not required
- Some experience, or a demonstrated interest in, working with non-technical users: explaining technical concepts in plain language, gathering requirements, or presenting findings.
- Final-year graduates with strong AI project portfolios are encouraged to apply.
Essential Technical Skills
- Proficiency in Python for AI/ML pipeline development and data processing.
- Working knowledge of LLM-based application development: RAG pipelines, prompt engineering, and structured output handling.
- Familiarity with agentic AI design patterns and orchestration frameworks – multi-agent architectures, tool-use, and human-in-the-loop workflows.
- Exposure to cloud-native development (Microsoft Azure preferred) and vector databases
- Good understanding of document processing pipelines — OCR, PDF/image extraction, chunking strategies, and multimodal content handling.
- Basic understanding of API design (REST/ async) and microservices integration.
Advantageous — Domain & Industry Knowledge
- Familiarity with maritime, offshore, or similarly regulated, asset-heavy industries.
- Exposure to compliance-driven environments where audit trails and regulatory alignment are operational necessities.
- Familiarity with SharePoint / Microsoft 365, SAP, or equivalent enterprise platforms.
Personal Attributes
- Curiosity: Genuinely interested in how operations work and how AI can improve them — not just in the technology itself.
- Structured thinking: Manages concurrent tasks, communicates priorities clearly, and keeps stakeholders informed.
- Credibility with users: Communicates clearly without jargon. Earns trust through listening and delivering solutions that fit how people actually work.
- Ownership mindset: Takes initiative, follows through, raises blockers early — and knows when to escalate rather than proceed alone.
- Comfort with ambiguity: Makes sound technical decisions when requirements are evolving. Moves forward pragmatically and documents assumptions.
- Engineering rigour: Builds things others can understand, maintain, and build on. Takes documentation and testing seriously.
