Raoof Altaher
In this section
On this page

Posts tagged coding-agents

A dark title card reading AI-Native SDLC with Ultrapowers in red-to-purple gradient lettering, between two glowing translucent spheres

AI-NATIVE-SDLC

AI-Native SDLC with Ultrapowers

A practical walkthrough of the AI-native SDLC using Ultrapowers: the six-stage mapping, one ticket from brief to reviewed pull request, skill pipelines, human gates and current limitations.

A dark title card reading The AI-Native SDLC: A Developer's Playbook in red-to-purple gradient lettering, between two glowing translucent spheres

AI-NATIVE-SDLC

The AI-Native SDLC: A Developer's Playbook

A practical guide to the AI-native software development lifecycle, grounded in Anthropic's six-stage model and focused on artifacts, verification, human decisions and operational feedback.

A dark title card reading Introducing Ultrapowers in red-to-purple gradient lettering, between two glowing translucent spheres

ULTRAPOWERS

Introducing Ultrapowers

A development method your coding agent follows. It asks before it codes, plans before it builds, tests first, reviews every task, and stops at your gates. One command takes a ticket to a pull request; you approve, you merge.

Karpathy AutoResearch architecture showing prepare.py as the fixed evaluation pipeline, train.py as the agent-editable training loop, and program.md as the operating instruction layer

RESEARCH

What Is AutoResearch?

AutoResearch is a small but powerful experiment loop where a coding agent edits training code, runs timed experiments, and keeps only the changes that improve a fixed metric.