Skip to content
RM

Setup Guides

Practical developer environments, tooling configurations, and agentic workflows curated for engineers and vibe coders.

IDE & Workflow
Beginner·8 min read

Cursor & AI Coding Setup: From Scratch to Production

The complete environment setup I use daily for fast, secure vibe coding: key extensions, .cursorrules conventions, model routing and indexing.

  • ✓Optimal model selection (Claude 3.7 Sonnet vs GPT-4.5 vs reasoning)
  • ✓Strict project rules to prevent AI hallucination and circular edits
  • ✓Git branch hygiene and atomic diff review techniques
Cursor.cursorrulesVibe Coding
Request guide update →
Agent Architecture
Intermediate·12 min read

Setting Up MCP (Model Context Protocol) with Local Tools

Connect your AI agents directly to databases, GitHub, terminal CLIs, and browser devtools with safe, sandboxed permissions.

  • ✓Configuring mcpServers JSON safely without leaking credentials
  • ✓Building a custom Postgres/Redis MCP bridge for database inspection
  • ✓Browser automation testing with Chrome DevTools MCP
MCPClaudeDevToolsAutomation
Request guide update →
Infrastructure
Intermediate·10 min read

Docker Sandbox for Running AI-Generated Code Safely

Isolate AI code runs, background workers, and persistent databases inside Docker containers bound strictly to localhost.

  • ✓Multi-container docker-compose setup for Postgres 16 & Redis 7
  • ✓Network isolation preventing untrusted outbound connections
  • ✓Fast database migrations and seed script automation
DockerPostgreSQLRedisSecurity
Request guide update →
Local AI
Beginner·7 min read

Running DeepSeek-R1 & Llama 3 Locally with Ollama

Run open-source reasoning models on Mac (Apple Silicon) or Linux with zero subscription costs and 100% offline privacy.

  • ✓Quantization guide: 4-bit vs 8-bit RAM requirements
  • ✓Integrating Ollama with VS Code and terminal coding assistants
  • ✓Setting up system prompts and context window limits
OllamaDeepSeekLocal AIPrivacy
Request guide update →
Developer Ops
Advanced·9 min read

Git Worktrees for Multi-Agent Parallel Coding

How to run multiple autonomous coding agents in separate parallel branches without duplicating node_modules or storage.

  • ✓Branch isolation patterns for simultaneous feature development
  • ✓Safe merging and conflict prevention when agents code in parallel
  • ✓Cleaning up transient worktrees and branch stashes cleanly
GitWorktreesMulti-AgentCLI
Request guide update →

Need a custom setup guide or stack review?

I help startups and developers optimize their AI coding workflows, .cursorrules, and agent infrastructure.