Built an automated pipeline for query optimization which allowed relevance experiments to be
described as code and deployed in half the time as the previous manual process.
Facilitated simple, declarative state transfer between airflow tasks by creating a python
package for persistent dataclasses backed by S3.
Created new granular search API using a vector search model for semantic similarity.
Improved observability of running models and scheduled tasks by unifying monitoring of all
deployed code within the search context into Grafana.
Senior Software Engineer, Technical Lead
Rewrote the shared navigation for all Pluralsight pages in Svelte, reducing the deployed
size by a factor of three and the dependency surface by 50%, making the code easier to
maintain going forward.
Rewrote infrastucture in AWS CDK, giving the full stack engineers on the team direct control
over their infrastructure.
Built the new search experience and relevance engine for the learner experience site.
Created an experiment platform to iteratively improve search result quality using live
feedback on query performance. Over it's lifetime this platform has doubled the search
clickthrough rate.
Translated between the the needs of search as a product and the highly technical search
domain.