A Building Agentic AI

The Blog

Writing on agentic AI.

Practical, honest pieces on building and designing agentic AI systems. Field notes from production work, written for engineers, architects, and technology leaders.

27 posts published 2 clusters · beginner + enterprise RSS →
  1. May 26 2026

    Giving Your Agent Memory: A Minimal Implementation

    "Memory" is one word for four different problems your agent has. The conversation buffer, summarization, episodic recall, semantic retrieval, and key-value preferences, each built from scratch in raw Python with no framework, plus the decision guide for which one you actually need.

  2. May 26 2026

    Tool Calling From First Principles (Before You Touch LangChain)

    Function calling, demystified. The under-the-hood mental model of how a model "calls a tool," a 40-line runnable example with no framework, the four things that go wrong in production, and when reaching for a framework actually helps.

  3. May 24 2026

    The Agent Loop, Explained: Perceive, Decide, Act, Observe

    The four-move loop that turns a language model into an agent. A diagram, a worked example, and the failure modes that come from breaking each step.

  4. May 24 2026

    Agentic AI Interview Questions: 30 Real Questions with Production Answers (2026)

    Thirty interview questions you actually get when applying for a senior agentic AI engineering role in 2026. Production answers from someone who has shipped these systems, not vendor talking points.

  5. May 24 2026

    Best Agentic AI Books in 2026: An Honest Engineer's Shortlist

    A short, opinionated reading list for software engineers learning agentic AI in 2026. Why so few agentic-AI-specific books exist yet, and which adjacent books actually hold up.

  6. May 24 2026

    Build an AI Agent From Scratch in Python, No Framework

    A tutorial that builds a real AI agent in about 150 lines of Python, with no LangChain, no LangGraph, no CrewAI. Just the agent loop, two tools, memory, and bounded autonomy.

  7. May 24 2026

    ReAct vs Plan-and-Execute: Which Agent Pattern, and How to Explain It in an Interview

    The two patterns most agentic interviews ask about. What each one actually does, when each one is right, and the hybrid most production systems end up using.

  8. May 23 2026

    What is agentic AI?

    A short, honest definition of agentic AI and how it actually differs from a chatbot or a generative AI app.

  9. May 22 2026

    RAG vs Agents: The Interview Question Candidates Get Wrong

    RAG and agents are not competing alternatives. They operate at different layers and compose. Why "RAG or an agent?" is a category error, the four mistakes candidates make answering it, what agentic RAG actually is, and the template for answering it in an interview.

  10. May 20 2026

    Agentic AI Glossary: 40 Terms, Plainly Defined

    Forty core agentic AI terms in plain language, grouped from foundations to production: the agent loop, tool calling, memory, design patterns, guardrails, MCP, and more. A reference you can scan or cite.