I lead the Media Analytics data science team in the Intelligent Conversation and Communications Cloud (IC3) group, which provides data-driven insights into call quality and reliability in Microsoft Teams and Skype.

As part of the AI/ML team in the Intelligent Conversation and Communications Cloud (IC3) group, I apply machine learning techniques to solve problems in the area of real-time communications.

As part of Skype's Call Quality team, I analysed telemetry and statistics from Skype and Skype for Business, investigating call quality issues and providing insights into where improvements can be made.This involved developing tools for automated analysis, performing in-depth investigations into specific call quality issues, and working with other teams to ensure end-to-end telemetry pipeline is working correctly.In addition, I provided expertise in developing tools for testing Skype in impaired network conditions.

As a software engineer in the Video Streaming team, I worked on developing and maintaining the protocols and testing infrastructure behind Skype video. My duties included determining test strategy, building frameworks to test my team’s products, and working with other video team engineers to instrument and analyse the performance of Skype’s video library. I also collaborated with other teams (e.g., Skype Labs, Skype Data team) to experiment with innovations from other parts of the company.In particular, I was involved in:- Developing and providing test automation for video-interoperability between Skype and Lync (Skype for Business);- Integrating and providing automated end-to-end testing for Skype's new telemetry pipeline;- Data analysis and performance evaluation studies (both in the lab and in the real world) of Skype video, including technical metrics and user-reported call quality data.

My research focused on understanding the performance impact of residential Internet links on streaming video.This involved measurement studies of RTP streaming to residential Internet users, analysis of the performance of these networks for video applications such as FEC, and developing statistical models to characterise the packet loss seen by residential Internet users.Specifically, I worked on the following:- Built a set of tools for active measurement of RTP video streaming.- Planned and executed a measurement study of RTP streaming to residential Internet users.- Developed statistical models for Internet packet loss.- Evaluated performance of forward-error correction for Internet streaming video.- Measured the properties of the Spotify peer-to-peer music streaming service.

Developing and maintaining simulation software (using Python and SimPy), to test the performance of “target tracking” algorithms.Working with embedded software for ZigBee sensor motes.