Tharindu Perera
Tharindu Perera is a software engineer and tech lead. He writes Refactix to share patterns from production work across AWS and AI engineering.
Connect on LinkedInArticles by Tharindu
23 articles, written between 2025-07-22 and 2026-05-26.
AI · 2026-04-22
Claude Code in CI/CD: headless patterns for real pipelines
Running Claude Code in CI/CD without the surprise bills. Headless flags, JSON output parsing, cost controls, and the failure modes nobody writes about.
AI · 2026-04-17
How to Use Claude Code Subagents: Parallel AI Work Without Chaos
Claude Code subagents run parallel AI work without context chaos. Config, dispatch patterns, model choice, and the mistakes that waste tokens, with examples.
AI · 2026-04-14
Claude Code Power User Guide: Skills, Hooks, Subagents
Configure Claude Code beyond defaults: skills, hooks, subagents, and CLAUDE.md memory. Working examples, team patterns, and common setup mistakes to avoid.
AI · 2025-11-18
LangGraph for autonomous AI agents: state machines that actually loop
How to use LangGraph for stateful AI agents that branch, loop, retry, and pause for human approval, with code for state, cycles, and checkpointing.
AI · 2025-11-13
Autonomous AI agents in CI/CD: building pipelines that reason
Build autonomous AI agents for CI/CD with LangGraph: code review, test generation, security analysis, and deployment decisions, with the operational patterns that keep them honest.
AI · 2025-10-10
Multimodal AI Development with GPT-4o and Gemini Vision
Working code for building multimodal apps with GPT-4o and Gemini, covering image analysis, document processing, visual debugging, and design-to-code.
AI · 2025-10-03
Production-Ready AI Agents with LangChain
How to build AI agents with LangChain that hold up in production: architecture, memory, tool integration, and deployment patterns, with working code.
AI · 2025-09-26
AI Coding Assistants Compared: Copilot, Cursor, and Claude
How GitHub Copilot, Cursor, and Claude actually compare for day-to-day coding work. Setup, productivity numbers, ROI math, and where each tool falls down.
AWS · 2025-09-19
Cutting AWS Lambda cold starts to under 100ms
How to get AWS Lambda cold starts under 100ms with SnapStart, smarter runtime choices, smaller packages, and provisioned concurrency without blowing the budget.