Deep dive into nuML's Descriptor pattern that bridges object-oriented class structures with mathematical ML algorithms. Includes both attribute-based markup and fluent interface approaches for late-binding scenarios.

Presenting the nuML machine learning library at CodeMash with a focus on accessibility and ease of use. Addresses predictive challenges, performance considerations, and provides talk materials for both supervised and unsupervised learning.

Visualizing how linear classifiers work by drawing separators through data points. Explores the concept of generalization and demonstrates the kernel trick for handling non-linearly separable data by projecting to higher dimensions.

Breaking down supervised learning into binary classification, multi-class classification, and regression using .NET data types. Shows how the Perceptron algorithm can learn from labeled examples and includes model serialization for reuse.

Introducing machine learning to .NET developers by removing the math barrier with an intuitive attribute-based approach. Demonstrates automatic feature selection using simple class decorators, making ML algorithms accessible without deep statistical knowledge.

Building an interactive scheduler control that allows users to select dates and add metadata. Demonstrates the commando-style approach of prototyping HTML, CSS, and JavaScript first before integrating with MVC.

Enhanced the JavaScript debugger to handle complex objects and arrays through recursive introspection. Demonstrates the power of the 'var' keyword in recursive functions and provides clean visualization of nested data structures.

Building highly interactive client-side controls requires effective debugging tools. Created a lightweight JavaScript debugger inspired by jQuery that works across all browsers, especially for IE environments where Firebug isn't available.