Technical notes on AI agents, cloud infrastructure, and development workflows.
Gemini CLI: Google's coding agent for the terminal
Teach Gemini CLI to write executable, single-file Python scripts with an inline dependency block, a technique that lets you run scripts without having to run pip install first.July
11, 2025
I started with a binary .FIT file from my sports watch and used
Gemini CLI to parse it, explore the data, and generate insightful plots about my run.July
10, 2025
Use npx to run the nightly and keep your stable gemini version intact while letting you test the latest features.July
09, 2025
This week, I used the Gemini CLI to generate scripts, and write code. Here are six practical tips that go beyond basic code generation.July 02, 2025
A look under the hood of the Gemini CLI to understand how it learns about a codebase, from the initial snapshot to on-demand learning with tools.July 02, 2025
Here are three useful keyboard shortcuts for the Gemini CLI that I have found to be particularly helpful.July 07, 2025
Google ADK: Google's open-source framework for creating AI agents
Learn how to extend the capabilities of your ADK agent by adding custom tools. This guide walks you through creating a Python function and integrating it as a tool, enabling your agent to interact with external systems and data.July 08, 2025
This guide shows how to quickly set up a local agent development environment, create a new agent, and run the ADK web UI.July 03, 2025
By transcribing and analyzing Gemini's podcast-style Audio Overviews, I uncovered a replicable five-phase framework that anyone can use to transform dense documents into engaging, conversational deep dives.July 01, 2025
Get ready for Cloud Next 2025, where Cloud Run takes center stage with sessions covering everything from new features and serverless GPUs to AI agents and high-availability services.April 07, 2025