Term & Causal Extraction
Automated reasoning methods for extracting domain-specific terms and causal relations from scientific text, combining symbolic ontologies with neural language models.
I'm a PhD student at the Technical University of Cluj-Napoca. My research focuses on term extraction, causal reasoning, and multi-agent platforms that pair structured knowledge with large language models. Applications span biomedical literature and ontologies, medical chatbots, battery informatics, and nickel-based superalloy design. I'm also a Teaching Assistant for Knowledge-Based Systems, Intelligent Systems, and Artificial Intelligence labs.
Automated reasoning methods for extracting domain-specific terms and causal relations from scientific text, combining symbolic ontologies with neural language models.
Agent-based pipelines that build and continuously update OWL ontologies from biomedical literature, with rule-validated extraction and DL-Learner axiom discovery.
Retrieval-augmented generation grounded in structured knowledge graphs, reducing hallucination in medical question-answering systems.
Practical adoption of large language models for student assessment generation in engineering education, examining context requirements, fairness, and evaluation.
MCP-orchestrated agent pipelines for tasks like automated disinformation detection, where reasoning needs to chain multiple specialized tools.
Multi-agent platforms grounded in structured knowledge graphs for expert-in-the-loop scientific workflows, including nickel-based superalloy design and other materials-informatics tasks.
Reconciling Manufacturer Claims with Measured Degradation in Commercial Lithium-Ion Cells: A Provenance-Aware Knowledge Graph with Coverage-Gated Abstention
Batteries Vol. 12(8), 314 · MDPI Paper ↗ Code ↗ Demo ↗
@article{lecu2026batterykg,
author = {Lecu, Alexandru and Hawizy, Lezan and Groza, Adrian},
title = {Reconciling Manufacturer Claims with Measured Degradation in Commercial Lithium-Ion Cells: A Provenance-Aware Knowledge Graph with Coverage-Gated Abstention},
journal = {Batteries},
volume = {12},
number = {8},
pages = {314},
year = {2026},
publisher = {MDPI},
doi = {10.3390/batteries12080314}
}
Engineering an ontology for Age-Related Macular Degeneration using Agentic AI and DL-learner
RuleML+RR 2026 20th International Rule Challenge · Vilnius · CEUR-WS Vol. 4286 Paper ↗ Venue ↗
@inproceedings{borzan2026amdontology,
author = {Borzan, Calina Annemary and Lecu, Alexandru and Muntean, George Adrian and Nicoara, Simona Delia and Groza, Adrian},
title = {Engineering an ontology for Age-Related Macular Degeneration using Agentic {AI} and {DL}-learner},
booktitle = {Companion Proceedings of the 10th International Joint Conference on Rules and Reasoning (RuleML+RR 2026)},
series = {CEUR Workshop Proceedings},
volume = {4286},
publisher = {CEUR-WS.org},
address = {Vilnius, Lithuania},
year = {2026},
url = {https://ceur-ws.org/Vol-4286/rc-paper-2.pdf}
}
AlloyGraph: A Knowledge-Graph-Grounded Multi-Agent System for Nickel-Based Superalloy Design
Preprint SSRN · DOI 10.2139/ssrn.7022006 SSRN ↗ Code ↗ Demo ↗
@misc{lecu2026alloygraph,
author = {Lecu, Alexandru and Hawizy, Lezan and Birosca, Soran and Groza, Adrian},
title = {{AlloyGraph}: A Knowledge-Graph-Grounded Multi-Agent System for Nickel-Based Superalloy Design},
howpublished = {SSRN preprint},
year = {2026},
doi = {10.2139/ssrn.7022006}
}
Memory-First Fact-Checking: A Knowledge-Graph-Grounded Multi-Agent System for Misinformation Detection
Preprint arXiv:2608.29617 arXiv ↗
@misc{petrenciuc2026memoryfirst,
author = {Petrenciuc, Amelia and Lecu, Alexandru and Groza, Adrian},
title = {Memory-First Fact-Checking: A Knowledge-Graph-Grounded Multi-Agent System for Misinformation Detection},
year = {2026},
eprint = {2608.29617},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2608.29617}
}
Reducing Hallucinations in Medical AI: A Knowledge Graph-Augmented Retrieval System for Evidence-Based Age-Related Macular Degeneration Information
IEEE Access Vol. 13, 210624–210639 arXiv:2502.11108 Paper ↗ Preprint ↗ Code ↗ Demo ↗
@article{lecu2025amdrag,
author = {Lecu, Alexandru and Groza, Adrian and Hawizy, Lezan},
title = {Reducing Hallucinations in Medical {AI}: A Knowledge Graph-Augmented Retrieval System for Evidence-Based Age-Related Macular Degeneration Information},
journal = {IEEE Access},
volume = {13},
pages = {210624--210639},
year = {2025},
doi = {10.1109/ACCESS.2025.3643370}
}
MCP-Orchestrated Multi-Agent System for Automated Disinformation Detection
SYNASC 2025 27th Int. Symposium on Symbolic and Numeric Algorithms for Scientific Computing arXiv:2508.10143 Paper ↗ Preprint ↗
@inproceedings{avram2025mcp,
author = {Avram, Alexandru-Andrei and Lecu, Alexandru and Groza, Adrian},
title = {{MCP}-Orchestrated Multi-Agent System for Automated Disinformation Detection},
booktitle = {2025 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC)},
pages = {333--341},
year = {2025},
publisher = {IEEE},
doi = {10.1109/SYNASC69064.2025.00051}
}
Leveraging LLMs for Student Assessment Generation in Engineering Education: a Study on Context Requirements and Adoption
EAEEIE 2025 34th Annual Conference, Cluj-Napoca · IEEE Paper ↗ Venue ↗ Code ↗
@inproceedings{lecu2025quizgen,
author = {Lecu, Alexandru and Ciupe, Aurelia and Groza, Adrian and Orza, Bogdan},
title = {Leveraging {LLMs} for Student Assessment Generation in Engineering Education: a Study on Context Requirements and Adoption},
booktitle = {2025 34th Annual Conference of the European Association for Education in Electrical and Information Engineering (EAEEIE)},
pages = {1--8},
year = {2025},
publisher = {IEEE},
doi = {10.1109/EAEEIE65428.2025.11136216}
}
Using LLMs and Ontologies to Extract Causal Relationships from Medical Abstracts
Procedia Computer Science Vol. 244, 443–452 · ACLing, Dubai · Elsevier Paper ↗ Code ↗
@article{lecu2024causal,
author = {Lecu, Alexandru and Groza, Adrian and Hawizy, Lezan},
title = {Using {LLMs} and Ontologies to Extract Causal Relationships from Medical Abstracts},
journal = {Procedia Computer Science},
volume = {244},
pages = {443--452},
year = {2024},
doi = {10.1016/j.procs.2024.10.219}
}
Live demo · Materials informatics
Multi-agent AI platform for nickel-based superalloy design. Companion to the SSRN preprint.
alloygraph.utcluj.ro ↗
Live demo · Battery informatics
Provenance-aware knowledge graph reconciling manufacturer claims with measured degradation in commercial Li-ion cells. Companion to the Batteries 2026 paper.
batterykg.utcluj.ro ↗
Live demo · Medical RAG chatbot
Knowledge-graph-augmented retrieval chatbot for evidence-based information on Age-Related Macular Degeneration. Companion to the IEEE Access 2025 paper.
llmkgraph.utcluj.ro ↗
Source code for individual projects lives on GitHub.
Ph.D. in Computer Science
Technical University of Cluj-Napoca
Automated Reasoning and Causality for Term Extraction · Advisor: Prof. Adrian Groza
M.Sc. in Artificial Intelligence
Technical University of Cluj-Napoca
Early Warning System for Predictive Maintenance in Optical Inspection
B.Sc. in Computer Science
Technical University of Cluj-Napoca
Steering Support in Electrical Power Steering
Data Scientist
Data Scientist / Software Engineer