Matt Warren takes a look at how to measure, what to measure and how get the best performance from .NET code, considering examples from the Roslyn codebase and StackOverflow (the product).
John Langford discusses how to use Vowpal Wabbit in and as a machine learning system including architecture, unique capabilities, and applications, applied to personalized news recommendation.
Charles Lamanna talks about the scale and architecture of Microsoft’s Azure Management Gateway and how Azure API’s are built for high availability and for data sovereignty.
Jevgenij Nekrasov discusses doing meta-programming in .NET, including writing a custom DSL.
Pushpraj Shukla discusses how Microsoft Bing predicts the future based on aggregate human behavior using one of the largest scale data sets, and recent progress in large scale deep learnt models.
David Staheli discusses approaches Microsoft is taking to plugin development, sharing experiences in reusing code across plugins for different IDEs, with demos of plugins in Eclipse, IntelliJ, and VS.
Andrea Magnorsky discusses active patterns, computation expressions, parsers, using type providers and more. These language features help make code simpler and easier to maintain.
Felienne Hermans explains how she used F# to determine if the game Quarto can end up in a tie or if there is always a winner. The technique used can be applied to scheduling and register allocation.
Phillip Trelford shows through live demos data structures that are orders of magnitude more performant than lists.
Alena Hall presents various machine learning algorithms available in Accord.NET - a framework for machine learning and scientific computing in .NET.