About

I'm Santosh. I write about AI agents, software engineering, and the decisions that make a system useful beyond its first demo.

This blog turns my learning notes and small experiments into guides you can follow. Each guide starts with a concrete problem, explains the moving parts, and gives you something to try. Diagrams show the architecture; code and failure cases help you check your understanding.

Start with AI agents

AI Agents Explained: Models, Workflows, and Harnesses introduces the model, tools, workflows, and execution machinery behind an agent. No coding experience is needed to begin.

It opens Harness Engineering: From Models to Reliable Agents, a series of sixteen standalone modules covering agent loops, context, retrieval, evaluation, deployment, and everyday applications. The series page lists the articles that are published.

How I approach the examples

• Show the boundary. A model proposing an action and an application executing it are separate events.
• Say what was tested. An offline replay, a live model call, and a deployed service establish different things. The guides label those differences.
• Keep sources close. Links to documentation and public learning resources sit alongside the concepts they support.
• Leave room for correction. If an example is unclear or a source changes, point it out in the article comments.

The examples are learning exercises. Fictional policies and sample data are identified in the text; they are not claims about a real business or a production deployment.

Get in touch

For technical questions or corrections, leave a comment on the relevant article so other readers can follow the discussion.

For professional connections or collaboration enquiries, find me on LinkedIn. My projects are on GitHub, and you can follow my writing on Hashnode.

Prefer a feed reader? Subscribe through RSS.