PhD researcher in trustworthy machine learning for low-resource and safety-critical domains.
University of Southampton, United Kingdom
I am a final-year PhD researcher in Computer Science at the University of Southampton, working on trustworthy machine learning for low-resource and safety-critical domains: Bayesian uncertainty quantification, explainability, and privacy-preserving deployment.
My research asks how models can know what they do not know, explain their reasoning, and be deployed responsibly where data is scarce and mistakes are costly. I am the primary researcher on a competitively awarded Alan Turing Institute grant (PICASO), contributed to the United Nations AI Trust and Safety Re-Imagination Programme, and won two University of Southampton Doctoral College Director's Awards. I am currently on the job market for research scientist and applied scientist roles in industry or academia.
Uncertainty & explainability
Bayesian deep Gaussian Processes that let models express calibrated confidence and explain their predictions in low-resource, high-stakes settings such as medicine and content moderation.
Privacy-preserving deployment
Federated prompt tuning (FLiP) that adapts language models across languages without centralising sensitive data, keeping trustworthy AI practical where data cannot be shared.
Knowledge representation (PICASO)
Probabilistic conceptual spaces that model vagueness, ambiguity, and confidence, improving how AI systems make sense of the world under uncertainty.
Won two University of Southampton Doctoral College Director's Awards, for Public Engagement & Outreach and Knowledge Exchange & Enterprise.
Presented Towards Explainable Hate Speech Detection in Roman Urdu at IJCNN 2026 (IEEE WCCI).
Awarded an Alan Turing Institute grant for PICASO: probabilistic conceptual spaces for sensemaking under uncertainty.
Led research team TrustWeave in the UN AI Trust and Safety Re-Imagination Programme.
Towards Explainable Hate Speech Detection in Roman Urdu
IJCNN 2026Lightweight Cross-Lingual Federated Prompt Tuning for Low-Resource LanguagesPDF↗
LREC 2026BAKER: Bayesian Kernel Uncertainty in Domain-Specific Document ModellingPDF↗
WSDM 2025Uncertainty Modelling in Under-Represented Languages with Bayesian Deep Gaussian ProcessesPDF↗
COLING 2025Would You Trust an AI Doctor? Building Reliable Medical PredictionsLINK↗
WISE 2024Detecting Cybercrimes in Accordance with Pakistani LawPDF↗
LREC-COLING 2024Comparing Prompt-Based and Standard Fine-Tuning for Urdu Text ClassificationPDF↗
EMNLP 2023Exploring Data Augmentation Strategies for Hate Speech Detection in Roman UrduPDF↗
LREC 2022