AI safety · uncertainty · reliable learning
Jamiu Tunde Suleiman
Ph.D. Researcher in AI Safety & Reliable Machine Learning
I am a Ph.D. student in Electronic and Electrical Engineering at Kyungpook National University in Daegu, South Korea. My research asks a practical question: how can learning systems recognize when their knowledge is insufficient, conflicting, or unreliable? I study uncertainty quantification, out-of-distribution generalization, reliable retrieval-augmented generation, and safety in federated and continual learning.
Research
Current directions
Uncertainty-aware AI
Treating uncertainty in LLM as a measurable, multi-signal construct so models can identify unreliable predictions and communicate epistemic limits.
- Uncertainty quantification
- Failure detection
- Selective prediction
Reliable language models
Improving factual grounding and personalization in language models, with particular attention to retrieval noise and evidence quality.
- Retrieval-augmented generation
- Rationale quality
- Personalization
Safe adaptive learning
Designing federated and continual learners that recognize client drift, conflicting evidence, distribution shift, and unsafe updates.
- Federated learning
- Continual learning
- Distribution shift
Selected publications
Full list ↗Enhancing the Reliability of Retrieval-Augmented Generation for Personalized Language Models via Rationale Quality Modeling
2026Manuscript under review
ACM Symposium on Document Engineering
LM4UC Workshop @ AAAI 2026
EMNLP 2024
Sustainability
Electronics
News
Recent updates
Our work, 'Revising RVL-CDIP: Quantifying Errors and Test-Train Overlap' was accepted at DocEng 2026
Released the preprint “From Scarcity to Efficiency,” studying data augmentation for low-resource African machine translation.
Our work on de-identifying sensitive information in document datasets appeared at EMNLP 2024.
Published Styleformer-ART, a study of ancient artifact character augmentation for sustainable knowledge preservation.
Joined Kyungpook National University in Daegu, South Korea, for graduate research in Electronic and Electrical Engineering.
Experience & education
Full CV ↗2025 — Present
Ph.D. Student & Graduate Researcher
Kyungpook National University
School of Electronic and Electrical Engineering · Daegu, South Korea
2023 — Present
Researcher
Machine Learning Collective (MLC)
Open machine learning research community
2023 — 2024
M.Sc. in Electronic and Electrical Engineering
Kyungpook National University
Daegu, South Korea
2021
Software Engineer
MyClinify
Nigeria
2013 — 2018
B.Eng. in Computer Engineering
Bayero University Kano
Upper Second-Class Honours · Kano, Nigeria
Get in touch
I welcome conversations about research collaborations, internships, and ideas in AI safety and uncertainty.
Shoot ↗