The Search Capability Program has built a heavily-tuned ElasticSearch-based search engine currently servicing millions of requests per day, serving up highly relevant search results for Business, Government and Points of Interest for Yellow Pages, WhereIs and numerous external partnerships.
Some of the activities I have conducted as part of the Search Capability team range from tasks in the following areas:
Information Retrieval:
continuously tuning and enhancing the search algorithm for maintaining or increasing search quality based on stakeholders’ requirements;
exploring new capabilities of ElasticSearch, while leading the migration of the search engine from its original implementation in Solr;
supporting the development of internal tools for measuring search quality.
Machine Learning:
developing reinforcement learning algorithms and decision trees for automatic classification of businesses;
analysing automatic hierarchical clustering of business verticals using k-means, agglomerative and divisive strategies.
Data Analytics:
extracting and interpreting logs (in some cases data crunching) to discover search trends or scenarios which demonstrate that suggested enhancements are working or need to be reviewed;
producing dashboards and data visualisations that continuously monitor live data like NPS and organic searches.
Quick Prototyping:
creating use cases and proof-of-concepts that help other business units within the company improve their offering;
exploring the emerging use of chatbots and dialogue assistants such as Amazon Echo, Google Home, Facebook Messenger, Slack, etc., with some of these systems already live, and others to be released soon.