Welcome to arefiva
Exploring how to build software systems where AI agents reason about problems, take action, and operate with meaningful autonomy. This is where I share what I am learning about agentic development, the patterns that work, and the disciplines required to build these systems well.
Latest from Agentic Development
-
May 3, 2026
Writing Specifications with the AES Skill
The Red, Green, Refactor With an AI in the Loop post introduced what I call an Autonomous Execution Specification (AES), and the post on mARCH,...
-
Apr 28, 2026
The AI Co-location Race
In high-frequency trading, the firms that win are not the ones with the best algorithm alone. They are the ones who figured out that the...
-
Apr 21, 2026
AIBC: Building a Training Ground for Autonomous Agents
Up until now, I have explored how to work more effectively with an autonomous agent: writing specifications with enough precision that the agent has something...
-
Apr 18, 2026
Acceptance Criteria That Actually Work
The quality of specification determines the quality of execution. This idea appears in the Red, Green, Refactor post, which framed it as the red phase...
-
Apr 17, 2026
mARCH: The Agentic Release Cycle Harness
Marching orders, automated.
-
Apr 16, 2026
Specification-First Development: The Mental Shift
There is a specific pull toward starting with code. Requirements arrive with enough shape to begin, the architecture is familiar, and writing code gives you...
-
Apr 15, 2026
Red, Green, Refactor With an AI in the Loop
Test-Driven Development (TDD) rests on a deceptively simple mantra: Red goes to Green goes to Refactor. Write a failing test, make it pass, then clean...
-
Apr 15, 2026
Welcome to Agentic Development
Welcome to the Agentic Development section of this blog. This is where I explore building software systems that use AI agents to reason about problems,...