Independent project · In progress
Food-production forecasting
Exploring Temporal Fusion Transformers for multivariate food-production modelling and interpretable temporal drivers.
TFT · Forecasting · AgricultureMeteorology × Data Science
Finding signal in complex systems.
I’m Jeffrey, a meteorologist and data scientist translating atmospheric observations, numerical models and machine-learning systems into forecasts and decisions that hold up in the real world.
Operational forecastingClimate MLRadar meteorology
Areas of expertise
01↗02↗03↗04↗05↗
Operational Meteorology & Nowcasting
Real-time public and aviation forecasting, severe-weather diagnosis and time-critical decision support.
Radar & Atmospheric Data Analytics
Extracting physical signals from radar, disdrometers, reanalysis and reproducible weather-data pipelines.
Machine Learning for Weather & Climate
Interpretable prediction across climate ensembles, environmental time series and spatiotemporal data.
Numerical Weather & Environmental Modelling
Production workflows for WRF, MPAS-A and CMAQ, from mesh design to forecast-field quality control.
Applied Meteorology & Climate Risk
Weather-sensitive decisions in energy, wildfire risk, agriculture and infrastructure operations.
Selected work
Climate ML · Imperial College London · 2025
Learning AMOC–NAO interactions from a large climate ensemble
EOF-based dimensionality reduction, lagged climate predictors, XGBoost and SHAP across a 40-member coupled-model ensemble.
Read case study ↗
Remote sensing ML · Imperial College London · 2025
Reconstructing radar-observed storms from satellite imagery
A U-Net trained on visible and infrared satellite channels to reconstruct storm-scale vertically integrated liquid fields.
Read case study ↗
Radar nowcasting · Hong Kong Observatory · 2023
One-minute phased-array radar for high-impact weather nowcasting
Rapid-volume, dual-polarisation analysis of hail, a waterspout and a severe airport-area squall thunderstorm.
Read case study ↗
Radar meteorology · Hong Kong Observatory · ERAD 2022
Raindrop microphysics across Southern China
Two-dimensional video disdrometer observations across three sites and multiple weather regimes, including tropical-cyclone rainbands.
Read case study ↗
Numerical modelling · ClusterTech · 2020
Variable-resolution forecasting of Typhoon Higos
A 128–8–4 km MPAS-A mesh used to examine the track and rapid intensification of a small tropical cyclone.
Read case study ↗
Physical meteorology · University of Reading · 2019
Tracing the moisture behind tropical-cyclone rainfall
Atmospheric moisture tracking over the Pearl and Yangtze River estuaries, with a focus on seasonality and ENSO.
Read case study ↗Supporting study
Predicting significant wave height from environmental conditions
A reproducible preprocessing and model-selection pipeline comparing linear regression, random forest and XGBoost across 68,751 observations.
Read supporting study ↗
Visualising rainfall, river and tidal flood risk
A team project integrating hydro-meteorological observations into interactive maps and visual comparisons for communicating potential flood risk.
Read supporting study ↗Developing work
Independent project · In progress
Weather-informed energy prices
Studying how meteorological variability, demand and supply signals interact in energy-price time series.
Energy · Time series · ExplainabilityOpen-source tool · In progress
metPipe: Northeast US
A reproducible pipeline for retrieving, preparing and analysing meteorological data for the northeastern United States.
SQL · Python · Data EngineeringResearch brief · Exploratory
Wildfire risk and mitigation
Structured research into utility wildfire-mitigation plans and operational fire-spread modelling, including Technosylva workflows.
Wildfire · Utilities · Decision supportBackground
Meteorological AI Training Expert
Micro1 · Remote
Building structured atmospheric analyses and HRRR-based reports for AI systems learning synoptic patterns and forecasting signals.
Scientific Officer
Hong Kong Observatory · Hong Kong
Radar operations, high-impact weather analysis, operational products and observing-system reliability in a real-time forecast environment.
Computational Scientist
ClusterTech · Hong Kong
Production HPC workflows and variable-resolution systems for weather, air quality and energy applications across Asia.
Education
MSc Environmental Data Science & Machine Learning
Imperial College London · High Merit
MSc Applied Meteorology
University of Reading · Distinction · Sutcliffe MSc Prize
BSc Earth System Science
University of Hong Kong · Second Class Honours, Division One
Working toolkit
MeteorologyOperational forecasting · Radar meteorology · Aviation weather · NWP · Observational and model-data analysis · Reanalysis
ProgrammingPython (PyTorch, TensorFlow, scikit-learn, xarray, pandas, NumPy, Matplotlib) · SQL · C++
Data & MLXGBoost · CNN · LSTM · Transformers · TFT · PCA/EOF · SHAP
SystemsHPC · Linux · Git · AWS · Azure · GCP