Knowledge Graph Scientist, ARTC

A*STAR RESEARCH ENTITIES; Singapore; On-site; SGD 120,000 - 180,000; Full time

Job summary

A*STAR Research Entities is seeking a highly skilled researcher to design and govern ontologies spanning EV battery, aerospace, and manufacturing domains. You will develop multi-hop reasoning, neuro-symbolic AI, and GraphRAG pipelines to extract and populate knowledge graphs from technical documents and sensor data.

The role requires a PhD with strong knowledge representation, and experience with LLM integration and cross-domain alignment.

Qualifications

• PhD in CS/AI with focus on knowledge representation, semantic web, or graph-based ML.

Responsibilities

• Design and govern the core ontology across parts, processes, regulations, and lifecycle states in EV battery, aerospace, and manufacturing domains.
• Develop and evaluate multi-hop reasoning methods, rule-based inference, and neuro-symbolic approaches for deriving new knowledge from graph facts.
• Research and prototype GraphRAG pipelines and LLM-to-KG extraction methods to populate the graph from technical docs, reports, and sensor logs.
• Develop cross-domain ontology alignment to unify terminology and schemas across data sources (engineering and supply chain).
• Research temporal knowledge graph structures to model how part conditions evolve across remanufacturing lifecycles for historical reasoning and predictive inference.
• Lead and contribute to research proposal development, technical roadmaps, and capability-building initiatives; identify new opportunities for the KG programme.
• Collaborate with the KG Engineer to operationalise research prototypes into production-grade graph systems.
• Prepare technical docs, research reports, and presentations for internal and external stakeholders.
• Publish findings at relevant research venues and domain conferences.

Skills

Python

Graph Databases

Ontology Design

SPARQL

Cypher

LLM Integration

Neuro-Symbolic AI

Temporal KG

Multi-hop Reasoning

NER

Relation Extraction

KG Completion

Education

PhD in CS/AI

Postdoc/Industry Experience

Job description

Roles and Responsibilities

• Design and govern the core ontology covering parts, processes, regulations, and lifecycle states in EV battery, aerospace, and manufacturing domains, ensuring methodology scalability regardless of the industry domain
• Develop and evaluate multi-hop reasoning methods, rule-based inference, and neuro-symbolic approaches that enable the system to derive new knowledge from existing graph facts
• Research and prototype GraphRAG pipelines and LLM-to-KG extraction methods that automatically populate the graph from technical documents, inspection reports, and sensor logs
• Develop cross-domain ontology alignment methods to unify terminology and schemas across data sources (e.g. engineering and supply chain), including handling conflicting industry standards
• Research temporal knowledge graph structures that model how part conditions and states evolve across the remanufacturing lifecycle, enabling historical reasoning and predictive inference
• Lead and contribute to research proposal development, technical roadmapping, and capability-building initiatives that align with industry needs and organisational strategy; proactively identify emerging technology opportunities and define new research directions for the KG programme.
• Collaborate with the KG Engineer to operationalise research prototypes into production-grade graph systems
• Prepare technical documentation, research reports, and presentations for internal and external stakeholders
• Publish findings at relevant research venues or domain-specific manufacturing and sustainability conferences

Job Requirements

Technical Skills

• Proficiency in Python, with experience in graph database systems (e.g., Neo4j, or Stardog)
• Strong foundation in knowledge representation and reasoning, including ontology design (OWL, RDFS, SKOS) and graph query languages (SPARQL, Cypher)
• Deep understanding of knowledge graph embedding methods (e.g., TransE, RotatE, ComplEx) and link prediction techniques
• Experience with LLM integration for knowledge tasks: Retrieval-Augmented Generation (RAG), information extraction, or fine-tuning language models on technical text
• Familiarity with neuro-symbolic AI approaches combining neural learning with logic-based reasoning
• Proficiency in one or more of the following areas:
  • Temporal knowledge graphs and evolving graph structures
  • Multi-hop reasoning and path-based inference
  • Named entity recognition and relation extraction from technical documents
  • Knowledge graph completion and missing link prediction

Bonus Skills (Preferred but not required)

• Familiarity with industrial or manufacturing data formats (BOMs, maintenance records, inspection schemas, ISO standards)
• Experience with digital twin frameworks or manufacturing informatics platforms

Qualifications

• PhD in Computer Science, Artificial Intelligence, or a related engineering field, with a focus on knowledge representation, semantic web, or graph-based machine learning
• Preferred: Demonstrated track record of independently securing competitive research grants, funding awards, or industry-sponsored projects.
• Preferred: At least 2 years of relevant post-doctoral or industry research experience, preferably with exposure to industrial or manufacturing knowledge domains

The above eligibility criteria are not exhaustive. A*ASTAR may include additional selection criteria based on its prevailing recruitment policies. These policies may be amended from time to time without notice. We regret that only shortlisted candidates will be notified.