The AI Agents Series
This series documents my learning journey with AI agents, informed primarily by hands-on experience using Codex across a range of projects.
The field is advancing—and the ways we work with agents are changing—so quickly that I often feel as though I am struggling to keep up. For me, writing this blog is a way to capture the experiences and observations I gather along the way. It helps me synthesize and share ideas I find useful, while also providing a constructive way to acknowledge and manage the anxiety that comes with such rapid change.
AI Agents 101
This section provides a foundational map of AI agents, covering the essential concepts needed to understand how they work. The initial drafts were created with Codex and will be expanded and refined over time as I gain new knowledge and practical experience through challenges at work.
Deep-Dive Topics
This section explores the topics that interest me most and that I want to study more deeply with Codex.
- Self-Improving Agents: Foundations and Advances
- Building an Agent Harness with Jev — a tutorial with screenshots on typed decisions for model routing, tool-call checks, and online evaluation.
- TODO: Deep Research Agents
- TODO: Memory Design in Codex and Claude Code
My Working Experience with AI Agents
This section captures my practical experience working with AI agents. It includes everything from concise tips and tricks to longer posts that bring together the skills and workflows I use in my day-to-day work.
- Working with Codex
- My Codex Skills as a Research Scientist
- Lessons from Engineers and Scientists — a collection of useful experiences from engineers and scientists working with AI agents.
Enjoy Reading This Article?
Here are some more articles you might like to read next: