Skip to main content
AI Do Now

Building scalable systems with AI - real workflows, patterns, and learnings

What You’ll Find Here

How I actually use AI agents to build production systems - not toy examples.Topics include:
  • Multi-agent coordination (Evaluator, Builder, Verifier)
  • Plan → Implement → Verify workflow
  • Auto-remediation and self-healing code
  • Prompt engineering for code quality
  • Token optimization strategies
AI isn’t magic. I share what breaks, why it breaks, and how to work around it.You’ll see:
  • Architecture decisions AI gets wrong
  • Security vulnerabilities AI misses
  • Novel problems AI can’t solve
  • How to know when to override AI suggestions
The best results come from humans and AI working together, not AI alone.Including:
  • When to use AI (boilerplate, patterns, tests)
  • When to use human judgment (architecture, security)
  • How to verify AI output effectively
  • Cost vs value trade-offs (token usage)
I’m building a production-grade multi-tenant SaaS platform as the vehicle for testing these AI workflows.The system includes:
  • Event sourcing with DynamoDB
  • Multi-tenant data isolation
  • Rust macro-driven development
  • Four-level testing strategy
But the system is the example, not the focus. The AI workflow is the story.

Start Exploring

Who Am I?

I’m a software engineer exploring how AI can accelerate building complex systems without sacrificing quality. I believe:
  • AI is a force multiplier - When used correctly, with humans in the loop
  • Workflows matter - Multi-agent patterns beat single-prompt coding
  • Verification is critical - AI output must be independently reviewed
  • Share the failures - AI mistakes teach us how to improve collaboration
Note: All content here is from my personal projects and learning journey. It does not represent my employer’s views or technologies.

Stay Connected


New content published regularly (articles weekly, video shorts twice per week)