Making data driven decisions through experimentation is an extremely important part of LinkedIn’ culture. It’s deeply ingrained in our development process and has always been a core part of Linkedin’s DNA. We experiment with everything from UI designs to back-end AI algorithms and infrastructure upgrades. It’s how we innovate, grow, and evolve our products to best serve our members and customers. It’s how we make our members happier, our business stronger, and our talent more productive.

We have built an end-to-end user experience and platform called T-REX, to enable targeting, ramping and experimentation, where a user can quickly set up an experiment against the targeted member group, ramp the new feature and quantify its impact in a scientific and controlled manner across all LinkedIn’s apps. Not only does T-REX allow easy design and deployment of experiments, but it also provides automatic analysis which is crucial in increasing adoption of A/B experiments. Being fast, reliable, and flexible, T-REX platform covers almost all the use case scenarios, including but not limited to LinkedIn.com, emails, AI relevance and ads. Every day, more than forty thousand experiments are run on nearly eight thousand metrics computed to accelerate innovations in every aspect of LinkedIn.

The team is constantly looking for ways to achieve better targeting, ramping, and experimentation. For example, there's a project underway where we intend to maximize the insights for an experiment with minimized time spent and a new feature that automates experiment ramping and multi-arm bandit adaptive experimentation. We also intend to move our infrastructure to the public cloud with Microsoft Azure and continue to develop a number of open source projects.

These challenges are extremely exciting and more importantly, are key to us delivering the best possible member experience.