Content by kinfey (19)

kinfey shows how to build a cloud-native evaluation harness for Azure AI Foundry skills using Foundry Hosted Agents, combining deterministic validators, an LLM judge that returns structured JSON, and a multi-turn adversarial attacker to catch regressions and compare models side by side.

kinfey explains how to run LLM agents that write and execute code without giving them a host-sized blast radius, using a MicroVM sandbox. The post walks through a real pipeline (a daily Mandarin World Cup podcast) built with Microsoft Agent Framework, Azure AI Foundry, and Hyperlight snapshot/restore isolation.

kinfey breaks down a cost- and security-aware blueprint for running a multi-agent SDLC “tower” on AKS, using AI Runway for in-cluster model serving, Kata MicroVM isolation for each agent pod, and MCP so GitHub Copilot Chat can orchestrate tools while keeping token spend predictable.

kinfey lays out a practical two-layer architecture for building and operating AI agents using Microsoft Agent Framework and Microsoft Foundry, with GitHub Copilot acting as a “coding agent” guided by versioned SKILL files. The post uses the ZavaShop workshop to show tools, MCP, workflows, evals, and deployment/ops guardrails across Python and .NET.

kinfey explains how AI Runway turns LLM deployment into a Kubernetes-native platform capability, using a unified ModelDeployment CRD to run and operate inference reliably across multiple engines and providers. The post walks through local CPU setups and an AKS-based path to production, with practical guidance on cost, observability, and governance.

kinfey introduces “AgenticOps” using the microsoft/AKS-Lab-GitHubCopilot workshop, showing how six scoped GitHub Copilot Custom Coding Agents plus the remote Copilot Coding Agent can author specs, code, tests, and deployments with explicit ownership and refusal rules, gated by evals and CI/CD.

kinfey explains why AI agents running model-generated code need stronger isolation than standard containers, then walks through deploying a GitHub Copilot SDK agent on AKS using Kata Containers (kata-vm-isolation) plus layered hardening like seccomp, NetworkPolicy egress allowlists, and deny-by-default tool permissions.

kinfey compares Anthropic’s “Managed Agents” architecture with Microsoft Foundry Hosted Agents, then walks through a Python sample that combines a stateless orchestrator, replaceable execution sandboxes, and an append-only event log, with Azure-native hosting, observability, and identity.

kinfey explores Phi-4-Reasoning-Vision-15B, Microsoft's new vision reasoning SLM. The article provides developers with detailed design analysis, example code, and real-world applications for visual understanding and actionable decision-making.

kinfey explains how to build a Kubernetes architecture for AI blog generation using dual sidecar containers: one for the GitHub Copilot SDK and another for skill management, offering actionable insights on modern DevOps and containerization strategies.

kinfey presents a detailed walkthrough for deploying OpenClaw—an open-source AI assistant—on Azure Windows 11 VMs, focusing on automation, security, and best practices for technical professionals.

In this in-depth guide, kinfey demonstrates how to combine the GitHub Copilot SDK, Agent-to-Agent Protocol, and Azure cloud-native practices to build scalable, collaborative multi-agent AI systems. Key technical takeaways and deployment patterns are provided.

kinfey delivers a practical exploration of hybrid AI system design, highlighting how Microsoft Foundry Local and GitHub Copilot SDK automate the conversion of GitHub README files into professional PowerPoint presentations. This article demystifies local and cloud AI model orchestration for real developer scenarios.

kinfey demonstrates building automated technical update tracking agents using the GitHub Copilot SDK and CLI, sharing a real-world case study on powering daily blog post generation for Microsoft's Agent Framework. The guide includes architecture, code samples, and best practices.

kinfey presents a comprehensive exploration of building an agentic podcast studio using the Microsoft Agent Framework, local SLMs, and VibeVoice. This guide reveals how edge-first AI orchestration empowers privacy, speed, and scalable creative automation.

kinfey presents an in-depth guide on building a local-first, multi-agent podcast studio using Microsoft's Agent Framework, complete with orchestration techniques, local LLM deployment, and advanced audio synthesis methods.

kinfey explains how to leverage Microsoft Foundry, Agent Framework, and VSCode for building powerful multi-agent AI solutions, detailing each step from visual design to code deployment.

kinfey explains how to set up and run your own AI language models locally using Foundry Local and Microsoft Olive, covering installation, model format conversion, file organization, and troubleshooting for beginners.

kinfey provides a hands-on guide to deploying OpenAI’s gpt-oss-20b and building intelligent agent applications locally using the AI Toolkit for Visual Studio Code, highlighting practical integration strategies for developers.