Development
Building Local AI Applications with .NET Aspire
Use .NET Aspire to orchestrate a Blazor UI, ASP.NET Core API, Qdrant, Redis, and optional Ollama model with health checks, telemetry, and tested local defaults.
27 min read
Technical articles
These articles turn hands-on engineering work into practical guidance: what problem a tool solves, how to implement it, and which trade-offs matter in production. Browse the latest writing below, or use Explore topics for a guided path through the archive.
Articles
AI
OpenRouter in .NET: Multi-Model Routing, Fallbacks, and Cost Controls
A friendly guide to choosing between AI subscriptions, direct APIs, and routing platforms. It then builds a well-behaved .NET 10 gateway with OpenRouter.
25 min read
Development
Introducing CLIF: Windows Desktop Automation for Scripts, Tests, and AI Agents
Meet CLIF, my open-source Windows automation CLI and local MCP server. Inspect WPF controls, script repeatable workflows, validate results, and give agents useful feedback.
16 min read
AI
Ollama Cloud vs Local vs Other Cloud Options for .NET Developers
An August 2026 snapshot for .NET developers comparing local Ollama, Ollama Cloud, and managed inference by limits, model choice, agent support, and measurable performance.
7 min read
Development
Evaluating AI Applications and Agents in .NET
Move beyond vibe checks with deterministic tests, golden datasets, LLM judges, and CI quality gates for AI applications built with .NET.
31 min read
Development
Building a C# MCP Server with a Blazor WebAssembly Inspector
Build a current, stateless Model Context Protocol server with the official C# SDK and pair it with a Blazor WebAssembly diagnostic inspector.
7 min read
Development
Building a Real-Time Blazor WASM AI Chat App with Microsoft.Extensions.AI
A practical .NET 10 guide to a guarded, streaming Blazor WebAssembly chat application built on Microsoft.Extensions.AI.
18 min read
AI
Running Qwen3.8 Flash-Next on Dual RTX 3060s with Proxmox and llama.cpp
An experimental case study of running Qwen3.8 Flash-Next on dual RTX 3060 GPUs and dual-socket Xeon hardware using Proxmox, NUMA interleaving, and an experimental llama.cpp build.
10 min read
AI
Building a Local AI Coding Workstation with Qwen3.8-27B and llama.cpp
An empirical case study of a dual-GPU Linux workstation for local coding assistance with Qwen3.8-27B, GGUF quantization, llama.cpp tuning, speculative decoding, and long-context benchmarks.
15 min read
AI
LiteLLM: Your Self-Hosted AI Gateway for Local and Cloud Models
Unify Azure AI Foundry, Ollama, LM Studio, and 100+ AI providers behind a single API with LiteLLM. Learn to deploy it locally with Docker Compose and PostgreSQL.
11 min read