All Posts

Three principles for building a practical software development harness: repo as system of record, auto-evaluation loops, and a dedicated black-box evaluator agent.

How I’m building a non-executive executive assistant with specialized AI agents for morning briefings, learning, tasks, blogging, and more.

A look at the file-based memory system that gives my AI assistant continuity across conversations.

A step-by-step guide to setting up OpenClaw, an open-source personal AI assistant on your own VPS with Telegram integration and practical automation examples.

How I built Leaf, a native macOS Markdown reader, using a spec-driven, vibe-coding loop with Codex CLI.

How we built a Cost-Quality Optimization system using Historical Replay and won the Portkey AI Builder Challenge hackathon.

How I built a customer feedback analysis system that saves days of manual research and creates synthesized reports for different audiences.

AI has exposed the reality of digital work - those who spent time on high-value deep thinking vs run-of-the-mill tasks. Time to aim 10X.

Build a complete AI coding environment accessible from your mobile phone in just 10 minutes for $6/month. Code anywhere, anytime using SSH, VPS, and AI coding agents.

Hands-on journey on fine-tuning Qwen3-0.6B on competitive programming problems using HuggingFace infrastructure and Claude Code.

To design memory systems for AI agents with a practical implementation of personalized quizzing app.

The no buzz-word primer on AI Agents you were looking for! Understanding agent primitives and build a GitHub PR Orchestrator Agent.

Building trustworthy AI workflows of your features.

Defining evaluation methodology, metrics & evals, and using LLM-as-a-judge for my product, a coding agent.

Learnings from building AI/LLM coding agents highlighting value of AI evals, trustworthy workflows and the impact created.

Drive more usage in Claude mobile app by improving critical user workflows

Strategy for winning in AI coding agents market

A framework for effective evaluation of AI powered products