cv

Teaching Assistant

University of Surrey

Teaching support across engineering systems, science principles and advanced pharmaceutical formulations.

• ENG2128 Engineering Systems and Management
• ENG1080 Science Principles for Engineers
• CHEM042 Advanced Pharmaceutical Formulations

PhD Researcher

University of Surrey

AI-enabled formulation design and experimental decision-making for dermal drug delivery.

• Develop probabilistic surrogate models, active-learning strategies and early-stopping algorithms for IVRT and IVPT experiments.
• Build interpretable dynamic models using domain-knowledge-constrained symbolic regression.
• Develop reproducible scientific LLM workflows for evidence extraction and dataset construction.

Technical Lead

Industrial data analytics projects

Applied process-data analytics to reservoir simulation and energy-market intelligence.

• Characterised dominant flow fields from numerical reservoir simulations using clustering and data-driven analysis.
• Built web-scraping and visual analytics components for a shale-gas trading market intelligence system.

Lead Researcher - Fault Diagnosis and Intelligent Early Warning

National Key R&D Programme of China project

Data-driven fault diagnosis for underwater production systems.

• Led development of intelligent early-warning algorithms in Python.
• Identified pump failures from operational process data.

PhD

University of Surrey

Chemical and Process Engineering

• Digital twin and AI methods for topical formulation design
• Supervisors: Professor Tao Chen and Dr Dimitrios Tsaoulidis

MEng

China University of Petroleum - Beijing

Control Science and Engineering, College of Artificial Intelligence

BEng

Panzhihua University

Automation, School of Intelligent Manufacturing

A modelling-based comparison of IVRT and synthetic-membrane permeation for early formulation screening

RSC Pharmaceutics

Quantifies alignment between IVRT and synthetic-membrane permeation assays.

Confidence-Bound Early Stopping of Experiments with Sequential Calibration

Chemical Engineering Research and Design, 230, 756-766

Early stopping with calibrated control of the false-stop rate.

Autonomous AI-Driven Design for Skin Product Formulations

Advanced Intelligent Discovery, e70100

A roadmap for assay-aware, closed-loop and increasingly autonomous skin formulation development.

Active learning-based adaptive optimisation for developing dermal drug formulations

Chemical Engineering Research and Design, 232

Adaptive batch allocation between expected-improvement and information-seeking formulation candidates.

An early decision-making algorithm for accelerating topical drug formulation optimisation

Computers & Chemical Engineering, 201, 109224

Probabilistic early decisions from partial IVPT trajectories.

Data augmentation using improved conditional GAN under extremely limited fault samples and its application in fault diagnosis of electric submersible pump

Journal of the Franklin Institute, 361(4), 106629

Virtual sample generation for fault diagnosis under extreme data scarcity.

Improving metric-based few-shot learning with dynamically scaled softmax loss

Image and Vision Computing, 140, 104860

A dynamically scaled loss for metric-based few-shot image classification.

Improved dynamic kernel PCA based on local preserving projections and its application for electric submersible pump fault diagnosis

The Canadian Journal of Chemical Engineering, 101(8), 4539-4554

Local-structure-preserving nonlinear dynamic process monitoring.

Green scheduling optimization of shuttle tanker route considering carbon emissions

Chemical Industry and Engineering, 39(2), 23-31

Multi-objective route scheduling with carbon-emission considerations.

Adaptive Batch Optimisation

Active learning for formulation optimisation under a limited per-batch experimental budget.

• Gaussian process regression
• Expected improvement
• Hypervolume contribution

Early Decision-Making Algorithm

Forecasting final permeation outcomes from partial IVPT trajectories.

• Uncertainty quantification
• Probability of exceedance
• Experimental resource saving

Domain-Constrained Symbolic Regression

Interpretable dynamic modelling and optimisation of dermal drug release.

• Domain constraints
• Dynamic modelling
• Interpretable optimisation

SkinMiner

LLM-powered extraction and verification of dermal formulation evidence.

• Scientific LLMs
• Evidence provenance
• Structured datasets

China Scholarship Council - University of Surrey Award

China Scholarship Council and University of Surrey

Doctoral scholarship, Grant No. 202306440060.

UK Turing Scheme Research Placement

UK Turing Scheme

Research placement at West China Hospital, Sichuan University.