Andrew Tabit

Software architect + AI coder 🥳 + AI agent builder 🤖

I help clients build cloud-native web and mobile apps under high load with Ruby, Node.js, and React. I also help developers transition to AI-augmented coding.

Atlanta, GA · LinkedIn

✅ Software Project Checklist

When I join a software project, aside from the obvious like CI/CD, DevOps, security, monitoring, and documentation, these are the practices I look for, establish, and support—and the red flags I work to eliminate.

✅ Well-written specifications
✅ Following standards and conventions
✅ Separation of concerns
✅ Small source files
✅ Readability over cleverness
✅ Correct variable/function naming
✅ Less is more
✅ Simplicity
✅ Small, focused commits
✅ Automated testing
✅ Code smell tools
✅ Time for refactoring
✅ Developers responsible for end product
✅ Using AI coding agents

❌ Vague or missing specifications
❌ Ignoring standards and reinventing the wheel
❌ Mixing concerns and responsibilities
❌ Large source files
❌ Clever code that's hard to understand
❌ Cryptic or misleading names
❌ Over-building and premature optimization
❌ Over-engineering
❌ Massive commits with mixed changes
❌ No tests or testing as an afterthought
❌ No code quality checks
❌ Rushing with no time for refactoring
❌ "It works for me locally" mentality
❌ Stick to manual coding

🌟 Technical Influencers

These people have helped shape my approach to software development. I align with their thinking on clean code, testing, developer experience, and AI-augmented development.

Kent Beck - TDD, Extreme Programming, simple design
Anders Hejlsberg - Delphi, C#, TypeScript, elegant language design
Martin Fowler - Refactoring, architecture patterns, clean code
Robert C. Martin (Uncle Bob) - Clean Code, SOLID principles, professionalism
DHH (David Heinemeier Hansson) - Ruby on Rails, simplicity, developer happiness
Yukihiro Matsumoto (Matz) - Ruby language, developer-centric design
Aaron Patterson (tenderlove) - Ruby/Rails core, performance optimization
Dan Abramov - React, Redux, teaching fundamentals
Kent C. Dodds - Testing best practices, practical JavaScript
Ryan Dahl - Node.js, Deno, runtime innovation
Charity Majors - Observability, developers owning production
Kelsey Hightower - Kubernetes, cloud-native, DevOps
Jessie Frazelle - Open source, security, developer tools
Simon Willison - AI coding tools, practical AI adoption

🔮 Future of Software Development

AI is transforming software development from writing code to orchestrating AI coding agents. Here's where I see this heading:

AI coding agents will write all code—we'll just describe what we need
Code review will become unnecessary as AI becomes proficient enough
Technology stacks will evolve: fewer programming languages, AI-optimized abstractions
Software will be dynamic and on-demand, with UI generated in real time
Applications will shift from static products to fluid, adaptive systems

In preparation for this bright future, developers must transition from code writers to AI orchestrators—manual coding is already obsolete!

I present these ideas at meetups and tech events, helping developers understand and prepare for this transformation.

🤖 My AI Coding Setup

I use Windsurf with Anthropic models and achieve 95% AI-generated code with high quality through a structured multi-agent workflow.

Setup:

Multi-agent team: Product Manager, Tech Lead, Developer
Three-level decomposition: Requirement → Task → Subtask
Developer Agent works on subtasks as the unit of work
I review and refactor immediately after each subtask

The key to quality is granular decomposition and immediate code review.

🚀 What Drives Me

Beyond building software, these things drive me:

Helping developers transition to AI coding before it's too late
Automating every business process at unprecedented scale with AI
Teaching non-technical people to build software through natural language
Creating AI-managed businesses where agents run operations autonomously
Advancing toward software entirely built by computers, not humans