I am a PhD researcher in Applied Mathematics at the University of Birmingham. My research focuses on temporal/time-series representation for foundation and world models, and on robust machine learning for non-stationary sequential data.

I study decomposition, retrieval, similarity, forecasting diagnostics, and symbolic structure as reusable representations for long-horizon temporal systems, including action histories, VLA trajectories, model rollouts, temporal memory, and variable-rich scientific or real-world time series.

• Temporal representation for foundation and world models: action histories, VLA trajectories, model rollouts, temporal memory, and long-horizon sequential behaviour.
• Time-series representation under non-stationarity: decomposition, stationarity-aware retrieval, forecasting diagnostics, classification, symbolic regression, and robust evaluation.
• Representation-driven research software: reusable decomposition and similarity tooling with inspectable temporal components, machine-readable outputs, and shareable reports.

• UKRI/AIRR compute resources (2026): two Gateway Project allocations on Isambard-AI: Time Series Language and Foundation Model and Language-Action Time-Series Tokenization for Efficient VLA Policies. Each project provides 10,000 GPUHR with nominal compute-resource value GBP 45,000; together they total 20,000 GPUHR with nominal compute-resource value GBP 90,000. Compute resources only, not direct cash funding.
• University of Birmingham College PhD Scholarship: full scholarship support for Applied Mathematics PhD study from Jan 2024 to Jul 2027, covering tuition fees and stipend/living costs.
• ICML 2026: first-author main-conference paper, Time-Series Decomposition as a Standalone Task: A Mechanism-Driven Diagnostic Benchmark.
• KDD 2026: co-author on Stationarity-Aware Retrieval-Augmented Time Series Forecasting.

• ICML 2026 | CORE/ICORE A*: Time-Series Decomposition as a Standalone Task: A Mechanism-Driven Diagnostic Benchmark. First author.
• KDD 2026 | CORE/ICORE A*: Stationarity-Aware Retrieval-Augmented Time Series Forecasting. Co-author.
• BIBM 2024 | CORE/ICORE B: iTARGET: Interpretable Tailored Age Regression for Grouped Epigenetic Traits. First author.
• Heliyon 2023 | JCR Q1: A novel online multi-task learning for COVID-19 multi-output spatio-temporal prediction. First author.
• ICONIP 2023 | Oral presentation | CORE/ICORE B: Correlated Online k-Nearest Neighbors Regressor Chain for Online Multi-output Regression. First author.
• ICONIP 2022 | Oral presentation | CORE/ICORE B: An Interpretable Multi-target Regression Method for Hierarchical Load Forecasting. First author.
• IJCNN 2020 | Oral presentation | CORE/ICORE B: A Novel Dynamically Adjusted Regressor Chain for Taxi Demand Prediction. First author.

Full publication list and CV: academic homepage / publications / CV.