Greater Melbourne Area
Geospatial Data Scientist & GIS Leader | Turning Earth Observation Into Environmental Impact I develop large-scale geospatial systems where cutting-edge research meets real-world impact. With a PhD in Earth observation from the University of Melbourne and 15+ years across research, government, and private sector, I specialise in translating advanced remote sensing and machine learning into production systems that solve critical environmental challenges. Current Role: At CSIRO, I lead the dynamic fuel science for Australia's National Bushfire Intelligence Capability serving emergency services across six states. I oversee continental-scale infrastructure processing 50+ TB of satellite data annually for 200+ users and deliver AI-driven geospatial solutions from research through operational deployment. I also lead MiniOzTree, a joint initiative between NASA, University of Copenhagen and CSIRO, Australia's first continental-scale individual tree detection pipeline processing 28,000+ LiDAR tiles and sub-metre satellite imagery, delivering high-resolution tree counting and classification at national scale. What I Do: I build systems that scale. My AI-driven vegetation classification models (92% accuracy) now run operationally across Australia's 7.7 million km². I have published 8 peer-reviewed papers while architecting enterprise geospatial systems integrated with business platforms serving hundreds of concurrent users. Technical Expertise: Remote Sensing: Multi-sensor satellite and airborne/UAV data fusion (Sentinel-2, Landsat, MODIS, Planet, multispectral UAV imagery), LiDAR processing, 30+ year time series analysis, vegetation monitoring, environmental forecasting Machine Learning: Deep learning (U-Net, ResNet, LSTM), TensorFlow, PyTorch, Scikit-learn, GANs, VAEs, random forest, HPC deployment Geospatial Tech: ArcGIS Enterprise, QGIS, Google Earth Engine, Python stack (GeoPandas, Rasterio, GDAL, xarray), PostgreSQL/PostGIS Cloud & Programming: AWS, Google Cloud, Azure, Python (advanced), R, SQL, PySpark, MATLAB Experience: Enterprise GIS implementation, land administration, cadastral mapping, team leadership, GIS-ERP-BI integration, WebGIS deployment, multi-million dollar programs Passionate about applying geospatial AI and earth observation at scale to solve critical environmental challenges — from bushfire intelligence to ecosystem restoration and carbon monitoring. 📚 8 publications | 🎓 PhD Earth Observation (The University of Melbourne) | 💼 Open to geospatial leadership & technical consulting
At CSIRO, I lead two flagship continental-scale geospatial AI programs that sit at the intersection of earth observation, machine learning and operational environmental intelligence. As Program Lead for Australia's National Bushfire Intelligence Capability (NBIC), I oversee continental-scale infrastructure processing 50+ TB of satellite data annually, delivering AI-driven fuel classification and vegetation monitoring solutions to 200+ users across six state emergency services. I also lead MiniOzTree — a joint initiative between NASA, University of Copenhagen and CSIRO — Australia's first continental-scale individual tree detection pipeline processing 28,000+ LiDAR tiles and sub-metre satellite imagery, delivering high-resolution tree counting and classification at national scale. My work involves building multivariate statistical and deep learning models trained on multi-source data including satellite time series (Sentinel-2, Landsat, MODIS, Planet), airborne and UAV multispectral imagery, LiDAR point clouds, Synthetic Aperture Radar, gridded weather data, ground surveys and real-time sensor feeds. Models are developed and deployed on AWS cloud infrastructure to ensure scalability and operational efficiency. Skills: Continental-scale geospatial AI, vegetation structure mapping, statistical and geospatial modelling, deep learning (U-Net, ResNet, LSTM), data classification and anomaly detection, satellite image processing, LiDAR processing, time series forecasting, cloud computing, process automation, technical documentation, peer-reviewed research Programming & Tools: Python, PySpark, R, SQL, PostgreSQL, ArcGIS Pro, QGIS, GDAL, Google Earth Engine, AWS, TensorFlow, PyTorch
My PhD research focused on satellite remote sensing of forest productivity and water status, with specific application to forecasting seasonal soil moisture and monitoring Eucalypt dryness across Victoria, Australia. Using 20+ years of NASA satellite sensor data, I developed advanced predictive models integrating tree physiology-based multispectral analysis, lagged correlation modelling, and machine learning and deep learning techniques. A key innovation was implementing SHAP (Shapley Additive Explanations) to enhance model interpretability, critical for real-world applications in natural resource management. This work laid the foundation for my subsequent research in continental-scale vegetation monitoring, carbon stock estimation and ecosystem health assessment.
Developed a machine learning-driven forest fuel moisture prediction model for the Victoria forest region, combining ground-level measurements from 40+ monitoring sites with satellite remote sensing data. The model provides accurate fuel moisture forecasts critical for bushfire risk assessment and emergency management decision making. This work directly informed subsequent operational tools developed for Australia's National Bushfire Intelligence Capability.
Led geospatial modelling and analytics for the global insurance and reinsurance sector, building catastrophe risk and agricultural forecasting products that fed directly into client underwriting and claims systems. • Managed a team of 4 GIS analysts delivering crop forecasting, drought prediction, and natural hazard modelling (hailstorm, earthquake, flood) for major insurance clients • Oversaw a project portfolio valued at $3M+ across agriculture, disaster risk, and climate domains, coordinating concurrent deliveries against strict client deadlines • Developed crop yield forecasting models integrated with client ERP systems for automated risk assessment and premium calculation • Designed and deployed automated geospatial pipelines on AWS, scaling to process petabyte-scale satellite and weather datasets • Presented analytics to insurance leadership, translating complex spatial models into business insights and risk metrics • Established QA protocols and standardised geospatial workflows across the organisation; led data acquisition strategy and vendor management
I was deputed at Mahalanobis National Crop Forecast Centre (MNCFC), here I worked simultaneously in two projects: NADAMS (National Agricultural Drought Assessment and Monitoring System): The project deals with Agricultural drought assessment at state and district level for various Indian states using MODIS, AWiFS and AVHRR satellite images. This includes extraction of indices like NDVI, NDWI, VCI, SMI etc. from the satellite images and incorporation of various external meteorological data to create a decision rule for drought declaration. FASAL(Forecasting Agricultural outputs using Space, Agro-meteorology and Land based observations): The project deals with multiple crop production forecast and yield prediction at national level using optical (LISS III and AWiFS) and microwave (RISAT-1) images. It includes supervised and unsupervised classification for optical images based on Ground truth points collected for the particular season, while hierarchal decision rule based classification for microwave images.
Here I worked on Delhi Public Geoportal. A WebGIS based Geoportal developed using Open Layers, Geoserver, PostgreSQL and Geoext libraries Skills developed: Geospatial Database Management, Geoserver Customization, Open layers, Command on PostgreSQL, Spatial Layer Styling using SLDs