At Coursera, I have been energized by development in production. This resolves

all bugs of the form 'cannot reproduce' because the development system and the production

system are one-and-the-same. Can this be done safely, where production, development and test are virtual systems layered on top of a single host platform?

Using Semantic Web, BigData and the cloud for omics (genomics etc).

My main responsibility is for the deep back end of the Syapse system, where the knowledge, about genomics, science and patients, is stored, managed and integrated.

I am eager that we manage the quality of our data sufficiently to extract actionable knowledge to improve cancer care both for individual patients, and other patients cared for by Syapse customers.

My initial responsibility at Syapse was the migration of our healthcare knowledge bases from a DIY system, to a commercial triple store (blazegraph from Systap), to which I also made open source contributions.

As the VP responsible for the development and delivery of TopBraid Live and TopBraid Ensemble,

I led a drive to quality improvements, through better processes, automated build, automated testing, continuous integration, more systematic manual testing and improved communication.

My role was very hands-on with substantial architectural and development responsibilities.

I helped initiate a move to use of scrum within the development team.

Managing the TopBraid product team, my tasks included: people management, release planning, defect tracking, liason with other functions (support, sales, marketing, training, CEO).

Key architectural decisions I have made are to make extensive use of: weak references, java.util.concurrent package and hierarchical deadlock prevention. The responsibilities of the two job-titles I had at TopQuadrant overlapped, and the promotion to VP of Engineering reflected what I was already doing.