Dev

Working in a ‘uv’ Workspace brings complication and overhead, but here I provide four real-life use cases that this mono-repo style makes more manageable.

I document my first trial run of a non-trivial agent with a multi-step task. The aim of the project is to develop an AI agent to assist in preparing for an academic journal club. In this initial run, I explore the DeepAgents library’s capabilities by making a simple request to download a journal article and convert it to Markdown format.

This is the beginning of what I hope will be a series of posts about using the DeepAgents library to build powerful AI agents. This posts describes how to get started running a simple research agent locally on a MacBook.

My thoughts after one week of thoroughly using Cursor AI coding environment on real projects at work.

This is just a quick post about a way I found in graduate school to quickly delete large amounts of data.

In persuit of the out-of-stock Colter Thorofare Net, I quickly created a GitHub Action to email me when it appeared back in-stock.

Spreading out computation is made simple in Python with the built-in multiprocessing module. Yet, it is not immediately obvious how to effectively portray the completion status in a progress bar. In this brief tutorial, I demonstrate how to easily and accurately display the progress of a multiprocessing pool.

In the course of my work, I stumbled across a misconception I had about the Final type annotation. I perceived it to mean that the annotated object should not be mutated. In this post, I demonstrate how this is an incorrect understanding of the Final type hint and a better solution to get help from the type checker in these situations.

A simple tutorial for getting up and running with LLMs running on your local computer.