Creating a strong embeddings model for code review data can empower users to build new solutions and enhance their capabilities. By making this model accessible to all, it can serve as a foundation for innovation, allowing individuals to push boundaries and advance further in their projects. This approach reflects a focus not only on data and experiences but also on infrastructure and core platform technologies, with the potential to revolutionize current practices and methodologies.
In this episode, Abi chats with Grant Jenks, Senior Staff SWE, Engineering Insights @ LinkedIn. They dive into LinkedIn's developer insights platform, iHub, and its backstory. The conversation covers qualitative versus quantitative metrics, sharing concerns about these terms and exploring their correlation. The episode wraps up with technical topics like winsorized means, thoughts on composite scores, and ways AI can benefit developer productivity teams.
(1:10) Insights in the productivity space
(7:13) LinkedIn's metrics platform, iHub
(12:52) Making metrics actionable
(15:35) Choosing the right and wrong metrics
(19:39) The difficulty of answering simple questions
(26:23) Top-down vs. bottom-up approach to metrics
(32:12) Winsorized mean and selecting measurements
(39:25) Using composite metrics
(46:57) Using AI in developer productivity