Research Engineer, Social & Cognitive Computing, IAIC

A*STAR RESEARCH ENTITIES; Singapore; On-site; SGD 90,000 - 130,000; Full time

Job summary

A*STAR RESEARCH ENTITIES is seeking a Knowledge Engineer to design, construct, and maintain knowledge graphs derived from diverse enterprise documents. You will leverage LLMs for knowledge extraction, ontology-driven structuring, and QA over graphs to support intelligent search and reasoning.

You will design schemas, model entities and relationships, normalize data, and implement storage in Neo4j/GraphDB. Collaboration with domain experts and stakeholders is essential to refine ontologies and

Qualifications

• Proficiency in ontology design and knowledge-graph modeling with semantic data representation.
• Strong experience with prompt engineering and LLM-based workflows.
• Hands-on work with LLMs (closed- and open-source).
• Experience with graph databases such as Neo4j and GraphDB.

Responsibilities

• Design, construct, and maintain knowledge graphs from heterogeneous documents.
• Leverage LLMs to extract structured knowledge from unstructured data.
• Develop LLM-powered QA and interaction pipelines over knowledge graphs.
• Collaborate with domain experts to refine ontologies and extraction logic.
• Process documents in multiple formats (Word, PDF, PowerPoint, Excel).
• Implement and manage knowledge graph storage using graph databases.

Skills

Ontology design

Knowledge graph modeling

Semantic data representation

Prompt engineering

LLMs

Tools

Python

Neo4j

GraphDB

Cypher

Job description

Job Purpose

The Knowledge Engineer will be responsible for designing, constructing, and maintaining knowledge graphs derived from heterogeneous enterprise documents. The role involves leveraging Large Language Models (LLMs) for knowledge extraction, ontology-driven structuring, and question answering over knowledge graphs to support intelligent search, reasoning, and decision-making systems.

Key Responsibilities

• Leverage LLMs to extract structured knowledge (entities, relations, attributes) from unstructured and semi-structured documents such as:Maintenance manualsProduct design documentsTroubleshooting guidesLessons learned reports
• Process documents in multiple formats, including Word, PDF, PowerPoint, and Excel.
• Design, construct, and maintain knowledge graphs, including:Schema and ontology designEntity and relationship modellingData normalisation and validation
• Implement and manage knowledge graph storage using graph databases such as Neo4j and GraphDB.
• Develop LLM-powered question answering and interaction pipelines over knowledge graphs, enabling:Natural language queriesContext-aware and explainable responses grounded in structured knowledge
• Collaborate with domain experts and stakeholders to refine ontologies, extraction logic, and use cases.

Required Skills & Experience

• Strong understanding of ontology design, knowledge graph modeling, and semantic data representation.
• Proficiency in Python, with experience building data processing and NLP pipelines.
• Solid understanding of prompt engineering and LLM-based workflows.
• Hands-on experience working with LLMs, including both:Closed-source models (e.g., GPT-based APIs)Open-source models (e.g., LLaMA, Qwen, etc.)
• Experience with graph databases, particularly Neo4j and/or GraphDB.

Preferred / Advantageous Skills

• Experience in information extraction, document understanding, or NLP pipelines.
• Familiarity with RDF, OWL, SPARQL, or Cypher is a plus.
• Experience integrating LLMs with structured knowledge sources (e.g., KG-augmented QA, RAG with graphs).

Key Competencies & Attributes

• Strong analytical and problem-solving skills, with the ability to translate unstructured information into well-defined data models.
• Ability to work independently while collaborating effectively within a multidisciplinary team.
• Good communication skills to engage with technical and non-technical stakeholders.
• Demonstrated willingness to continuously learn and keep up with advances in LLMs, NLP, and knowledge engineering.