AI and Knowledge Graph Engineer

AI and Knowledge Graph Engineer

Location: Manchester, UK
Contract Type: Contract Position

About the Project

Join a specialist software development team delivering a new, business-critical technology platform for an established international organisation. This greenfield development project involves modern cloud architecture, data-intensive applications, and AI-enabled capabilities. You will work as part of a multidisciplinary team alongside experienced software, AI, infrastructure, and design professionals, with direct involvement in taking the platform from development through to production. Security, scalability, maintainability, data protection, and production readiness are key priorities throughout the project.

Role Purpose

As the AI and Knowledge Graph Engineer, you will design and build the platform's knowledge representation and the AI capabilities that sit on top of it. This includes ontology and schema design, entity resolution, retrieval, evaluation, and ensuring robust production behaviour of AI-assisted features. Applicants must have delivered AI features into production and measured their impact, not only prototyped them.

Main Responsibilities

Knowledge Graph Design

• Design ontology and schema: entities, relationships, attributes, and the rules that constrain them.
• Model provenance so every assertion can be traced to its source document, extraction run, and confidence.
• Plan schema evolution and migration, keeping historical data queryable as the model changes.
• Define identifiers and resolution rules to prevent duplication of real-world entities across sources.

Ingestion and Entity Resolution

• Build pipelines to extract entities and relationships from structured and unstructured sources.
• Implement entity resolution with explicit precision and recall targets, including a review path for uncertain matches.
• Handle conflicting sources deterministically, recording which source prevailed and why.
• Validate ingested data against the schema and quarantine failures rather than admitting them silently.

Retrieval and AI Features

• Build retrieval over the graph and document content, combining structural and semantic search.
• Design prompts and tool interfaces to ensure model answers are grounded in retrieved evidence and properly cited.
• Set explicit behaviour for missing evidence, conflicting evidence, and low confidence, including refusal to answer.
• Control cost and latency through caching, batching, and model selection.

Evaluation and Quality

• Build evaluation sets from real tasks and measure accuracy, grounding, and refusal behaviour before release.
• Detect regression between model or prompt versions automatically and block promotion if regression is found.
• Monitor live quality and feed failures back into the evaluation set.
• Report honestly on system limitations and make these visible in the product.

Collaboration and Hand-off

• Work with UI/UX Designers to ensure evidence, confidence, and human decisions are clearly distinguishable in the interface.
• Provide clear interfaces and documentation for application engineers consuming these capabilities.
• Review technical constraints with the Technical Lead and agree on priorities ahead of each sprint.

Requirements

• Substantial professional experience building AI or data-intensive systems that reached production.
• Practical experience with knowledge graph or semantic data modelling, including ontology design and entity resolution.
• Strong Python skills, and competence with a graph database such as Neo4j and with a vector store.
• Experience with retrieval augmented generation, including grounding, citation, and refusal behaviour.
• Demonstrated evaluation practice: building evaluation sets, measuring quality, and preventing regression.
• Sound data engineering fundamentals: pipelines, schema validation, idempotency, and backfill.
• Able to explain model behaviour and its limits to non-specialists without overstating certainty.
• Comfortable working independently within a multidisciplinary team.

Desirable Experience

• Experience with SPARQL, RDF, or property graph query optimisation at scale.
• Experience fine-tuning or distilling models for a narrow production task.
• Familiarity with information extraction from scanned or semi-structured documents.