Cassie Breviu
Software Engineer & Product Leader · Windows @ Microsoft
Building at the intersection of AI and product for over a decade. I currently lead engineering and product projects in the Windows org at Microsoft. Previously I built end-to-end ML solutions across startups and enterprise, from PyTorch model authoring to production inference at scale. Passionate about turning research into real products people actually use.
Reframing How You See Engineering Value: Our Evolving Mindset as Developers Using Agentic Engineering Tools
There's a lot of noise about what engineering value means in the age of agentic tools. The journey from excitement to fear to real engagement reveals something important: your value was never in the code you type it's in the judgment, systems thinking, and experience behind it.
Concepts for Orchestrating the Agentic Future
The future of computing is rapidly evolving toward an agentic paradigm where AI-powered systems work collaboratively to enhance human productivity and creativity. This transformation requires a comprehensive approach that seamlessly integrates diverse tools, applications, and agents into a unified ecosystem. The orchestration of this agentic future isn't just about implementing LLM Models, it's about creating an intelligent infrastructure that connects everything we already have while paving the way for what's next in a scalable way.
Microsoft Build Session: Code-First LLMOps from prototype to production with GenAI tools | BRK110
Learn how to get from prototype to production with LLMOps and new developer tools for iterative debugging, evaluation, deployment, and monitoring. We’ll test and troubleshoot a copilot app using the code-first prompt flow SDK and frameworks like LangChain, interactively debugging an LLM chat application with VS Code. Then we’ll make sure it meets our thresholds for performance and safety, ship it, and ensure we can collect critical signals in production.
Empowering Your Business with AI: Introducing AI Application Templates
AI project starter templates for common GenAI scenarios. What is an AI project starter template anyway? These are GitHub repositories that have end-to-end solutions to jumpstart your AI endeavors. These templates serve as a foundation, allowing companies to harness AI benefits without reinventing the wheel.
GenAI RAG Chat Application
This sample has the full end-to-end process of creating RAG application with Prompty. It includes GPT 3.5 Turbo LLM application code, evaluations, deployment automation with AZD CLI, GitHub Actions for evaluation and deployment, an monitoring for production solutions.
AI Image Generation: Stable Diffusion in C#
This is the first and only all C# implemenation of the Stable Diffusion Image creation model.
OpenAI Whisper AI Model Implemented in C#
This project demonstrates the use of Olive to optimize an AI model for specific hardware, utilizing the Whisper Tiny English version as a case study. The optimized model is then deployed to the device and using ONNX Runtime we can execute both local and cloud-based inference.
Unreal Engine AI Real Time Style Transfer on Device
Apply a Style Transfer Neural Network in real time with Unreal Engine 5 leveraging ONNX Runtime.
Co-Authored of PyTorch: Learn the Basics
Co-Authored the Learn the Basics tutoral for creating deep learning models with PyTorch on Pytorch.org.