Back to blog

How I Use AI Agents in My Daily Engineering Workflow

From scaffolding projects to reviewing architecture decisions — how tools like Claude Code are changing the way I write software.

AI Claude Code Productivity Software Engineering

As a backend engineer working across public sector, healthcare, and financial services, I’ve been integrating AI agents into my daily workflow. Here’s how it’s changed the way I work.

The Shift

For years, software development has followed the same loop: read requirements, design, implement, test, deploy. AI agents don’t replace this loop — they accelerate every step of it.

I started using Claude Code as an AI pair programmer, and quickly realized it’s not just autocomplete on steroids. It’s a genuine collaborator that can:

  • Scaffold entire project structures in minutes
  • Write and review tests with context about the full codebase
  • Help design API schemas and database models
  • Debug complex issues by analyzing stack traces and logs
  • Generate documentation that actually reflects the code

Real-World Examples

Architecture Planning

When starting a new microservice, I describe the business requirements and constraints to Claude Code. It proposes an architecture, suggests technology choices, and identifies potential pitfalls — all before I write a single line of code.

Migration Projects

One of my key responsibilities at ti&m is migrating applications from Java to Kotlin and from Spring Boot to Quarkus. AI agents help by analyzing existing code patterns, suggesting idiomatic Kotlin replacements, and ensuring nothing breaks in the process.

Process Digitalization

Working on government platforms like dLIS (aviation licensing), I use AI to help model complex business processes with Camunda BPM. The agent understands workflow patterns and can suggest optimal process designs.

What I’ve Learned

  1. Prompt engineering is a real skill. The better you articulate your intent, the better the output.
  2. AI augments, it doesn’t replace. You still need deep domain knowledge to evaluate and refine what the AI produces.
  3. The feedback loop is key. Iterating with an AI agent on a design is much faster than whiteboarding alone.

Looking Ahead

AI-augmented development is becoming a core competency. Engineers who learn to work effectively with these tools will ship faster and build better systems. It’s not about replacing developers — it’s about amplifying what we can do.