Short Title: The .NET ML Engine: Designing a Native Machine Learning Pipeline Framework for Managed Applications Abstract: We introduce .NET ML Engine, a high-performance machine learning framework built natively for the .NET ecosystem, enabling developers to author, train, and deploy complete ML pipelines directly within C# and F# applications. Unlike Python-centric frameworks, our system introduces DataView—a relational-inspired data abstraction that supports lazy evaluation, immutability, and streaming access to datasets exceeding memory limits. The engine supports over 80 featureizers and 40 learning algorithms, all accessible through a unified .NET API, and provides seamless interoperability with Python via NimbusML, a Scikit-learn-compatible layer. In performance benchmarks against Scikit-learn and H2O across real-world datasets (Criteo, Amazon Reviews, Flight Delays), .NET ML Engine demonstrates up to 10× faster training times and superior scalability while maintaining competitive accuracy. The framework is production-hardened, having been deployed across hundreds of Microsoft services and open-sourced under MIT license, offering .NET developers a first-class, end-to-end ML experience without leaving their native stack.