|contract type:||consulting / remote|
|role:||Senior Search Engineer|
|time period:||Dec 2012 – May 2015|
search engines • data science • apache solr • column-store databases • amazon redshift • cloud infrastructure • amazon ec2 • amazon s3 • amazon emr • Python • Linux • numpy • scipy • matplotlib • ipython • scikit-learn
At Juju, I was the sole person responsible for conducting clickrate optimization based on query logs and related data. I conducted a series of one-off model-fitting projects, and was also responsible for the design, implementation, and continual improvement of the underlying infrastructure.
Juju is one of only a handful of job search engines and ad-exchange platforms for job-postings in the U.S. Given its cost-per-click revenue structure, clickrate optimization is of great importance to Juju.
The greatest technical challenge was the sheer scale of the dataset, with the core analytical table having well over a hundred attributes and hundreds of billions of records.
Throughout my time at Juju, I worked closely with the CEO, who often sought out my contributions, even on non-technical subjects. At one point, more than half of the Juju technical staff consisted of people interviewed and selected by me.