All Articles
10deep dives on AI, Agentic systems, LLMs, RAG & production.Your AI Agent Has Goldfish Brain — Two Ways to Fix It in Microsoft Agent Framework
A practical comparison of two memory architectures for Microsoft Agent Framework 1.0 — Mem0 as a tool the agent calls, and Azure AI Foundry Memory as a transparent context provider.
Plan It. Review It. Execute It: Building a Human-in-the-Loop Agent with LangGraph
A practical architecture for AI agents that pause for human approval, support replanning, and execute safely with checkpointed state.
Stop Making Users Wait: Why Your API Needs Background Workers
How to move long-running API work to asynchronous workers with Celery so user-facing endpoints stay fast, resilient, and scalable.
Running Claude Code with a Local LLM — Step-by-Step Guide
A complete walkthrough for connecting Claude Code to a locally-running LLM using Ollama, enabling offline coding assistance without API costs.
Why One Brain Isn't Enough: The Power of the Multi-LLM Chain-of-Debate
Single LLMs hallucinate, show bias, and miss perspectives. Chain-of-Debate orchestrates multiple models to critique and improve each other's outputs — producing answers that no single model could reach alone.
Configuring Claude Code with Azure AI Foundry: A Practical Step-by-Step Guide
A practical walkthrough to connect Claude Code to Azure AI Foundry using API keys and environment variables, then verify it inside VS Code.
From Vibe-Coded to Production: The Engineering Reality of AI Agents
The gap between a demo-quality AI agent and one that ships to production is enormous. Here's what actually breaks — and how to fix it.
Building a Tweet Critique Agent with LangGraph Reflection Pattern
A compact agent loop that separates generation from critique, iteratively improves tweets, and stops safely using a termination heuristic.
From Chaos to Clarity: How the User-Driven Workflow Planner Makes Agentic AI Design Effortless
Agentic AI systems fail when users can't express what they want clearly. The User-Driven Workflow Planner pattern lets users define agent workflows in natural language — no code required.