Experience to help process large datasets in enterprise databases and cloud storage. Data can be sourced as time-series records, grammar text or geo-locations. Object stores & key-index databases allow for fast lookup of large unstructured data, whereas relational storage, possibly with stored-procedures, allows normalised & structured data to be indexed & processed, consistent with the schema.

(Distortion Model for Polarimetric Calibration)
Download document

Offering the means to derive meaningful measures and comparisons from large data. Information may be accumulated in different formats and dimensionality but require consolidation to datasets showing the meaningful content.

(Statistical Models for Polarimetric SAR Data)
Download document

Analyzing connections and relationships between datasets allows measures and ranking to be applied to the extracted entities. However visualization and comparison requires choice of attributes & coordinate-systems. Experience in capturing and consolidating the relevant features of data.

(Spectrum Estimation of Multichannel ARMA models)
Download document

Supervised training & classification of speech to control applications on mobile & home devices. JavaScript-based frontend models used to train backend classifiers. Application-specific datasets from prototype trials. Scalability with international cloud deployment