The role of the software engineer is shifting from execution to orchestration, and it's happening faster than most of us realize. Dennis Vink, Principal Consultant at Xebia, breaks down how he approaches code modernization with AI, why fundamentals and system design matter more now than ever, and what the engineering role is actually becoming.

In this episode, we cover:

• Why you need to mature your old codebase before you can migrate away from it
• How to prove feature parity between legacy and modern systems
• Why vibe coding without architecture knowledge gives you zero control
• The shift from execution-focused engineering to orchestration
• Why Dennis worries about the next generation of engineers

Whether you're sitting on legacy code at work or wondering how your role as an engineer is evolving, this conversation will make you think about where you need to invest your time next.

Timestamps:
00:00:00 - Intro
00:00:51 - Dennis's Early AI Engineering Assignments
00:02:23 - Side Projects: Reviving a 20-Year-Old Game in Rust
00:04:36 - Why Vibe Coding Without Fundamentals Fails
00:05:15 - The Fundamentals You Need for Code Migration
00:06:45 - Proving Feature Parity with Automated Testing
00:08:12 - Writing Tests First as Risk Mitigation
00:10:13 - How Much Should You Care About Code Structure?
00:11:18 - Migrating in Small Pieces of Value
00:12:26 - Will Engineers Still Find Fulfillment in Building?
00:14:01 - How to Actually Start Side Projects (ADHD Brain)
00:15:34 - Why Pivoting Is No Longer Painful
00:16:12 - Prompting as the New Bottleneck
00:17:23 - Parallelizing Work Across Projects
00:19:08 - Why System Design Is the #1 Audience Demand
00:20:19 - AI as a Differentiator for Strong Architects
00:21:11 - Why the New Generation Should Worry
00:23:01 - Are Bootcamps Still Worth It?
00:25:15 - The Shift from Collaboration to Business Understanding
00:27:56 - Infrastructure as a Core Competency Bet
00:30:15 - Deterministic vs Non-Deterministic Code Generation
00:32:16 - Can This Approach Scale to Million-Line Codebases?
00:34:20 - Why a Finger-Snap Migration Would Scare You
00:37:01 - Where to Start with Your Own Legacy Codebase
00:38:43 - Which Languages Do AI Models Struggle With?
00:40:24 - Building Around Hallucination with Scaffolding
00:42:30 - Spec-Driven Development as the Future Way of Working
00:43:30 - Turning a Non-Technical Colleague into a "Developer" in an Hour
00:46:21 - When the House Is on Fire, That's When You Need Real Engineers

Projects we discussed:
Agent designer - hurozo.com
Game project - Zorlore.com (https://github.com/zorlore/)
Vibe coded solar system simulation - spacehaste.com

#SoftwareEngineering #SystemDesign #AIEngineering

Most senior engineers don't realize they're stuck until it's too late. The longer you stay, the more people around you have already decided who you are and what you're for. Ian Miell, CTO at Container Solutions, breaks down why this happens and how understanding the system around you is the first step to growing beyond it.

In this episode, we cover:

• Why staying too long gets you put in a box (and how to escape it)
• How your software architecture is shaped by money flows
• The 30% rule: why you should feel uncomfortable at work and what it means if you don't
• How to pitch to senior leadership and actually get buy-in
• Why AI makes distribution the real challenge, not building

If you're a senior engineer trying to grow beyond your current ceiling, this one is worth your time.

Timestamps:

00:00:00 - Intro

00:00:42 - How to Pitch to Senior Leadership and Get Buy-In

00:03:26 - Why You Should Feel Uncomfortable 30% of the Time

00:06:33 - How to Break Through a Seniority Ceiling

00:08:24 - The Burden of Context: Why Being the Go-To Person Traps You

00:10:16 - How Ian Became CTO Without Trying To

00:13:40 - Why a CTO's Job Is Mostly Coaching Now

00:18:20 - Understanding Incentives: The Key to Navigating Any Org

00:23:08 - Startups vs. Large Companies: Completely Different Rules

00:25:00 - Why AI Makes Distribution the Real Problem, Not Building

00:28:16 - The Hidden Maintenance Risk of Vibe-Coded Software

00:30:13 - Security and Compliance: More Nuanced Than Engineers Think

00:36:54 - Where "Architecture Follows the Money" Came From

00:42:36 - The Wrong Number of Customers: A Systems Thinking Story

00:47:23 - Why Engineers Think Individually Instead of Systemically

00:51:53 - How to Start Thinking in Systems

00:57:50 - How to Create Cross-Pollination in Consulting Teams

00:59:39 - What CTOs Actually Look for When Hiring

01:00:34 - Outro

#softwareengineering #systemsthinking #careergrowth

Most architects stop coding... and that's exactly where they lose their edge. Dennis Doomen has been a hands-on coding architect for 30 years, and his take is blunt: if you're not in the code, you can't make good architectural decisions. Period.

In this episode, we get into the real causes of codebase rot, why dogmatic pattern-following destroys teams, how Dennis uses AI tools to build open source projects without compromising his standards, and why documentation and decision records might be the most underrated investment a software team can make.

This one is for software engineers and architects who want to stay sharp, stay relevant, and build systems that actually last.

00:00:00 - Intro

00:01:05 - Why Dennis Refuses to Stop Coding (After 30 Years)

00:02:54 - The Only Way to Be an Effective Software Architect

00:04:43 - What Happens When Teams Copy Patterns Without Understanding Them

00:06:23 - Software Engineering Is About Battling Complexity

00:08:20 - When to Break Consistency to Reduce Complexity

00:09:24 - The Problem with Overzealous SOLID Principles

00:11:06 - The Future Where We Don't Care About Code Anymore

00:12:07 - How Dennis Built an Open Source Library with GitHub Copilot

00:14:18 - Accepting AI-Generated Code That Doesn't Meet Your Standards

00:16:39 - How to Use AI Without Losing Code Quality

00:17:41 - The Execution Is Accelerating — What Actually Matters Now

00:20:19 - Why Tests Are Your Safety Net in an AI-First World

00:23:44 - Lessons Learned from Letting AI Run Unsupervised

00:26:46 - Should Teams Standardize Which AI Tool They Use?

00:27:32 - Junior Devs and AI: Learning Skills vs. Speed

00:29:21 - How to Stay Curious and Critical in an AI-Assisted Team

00:33:43 - How to Build a Software Engineer from Scratch Today

00:34:38 - Dennis's Emoji-Based Pull Request Review System

00:36:45 - What AI Still Can't Do: Holistic Architectural Thinking

00:38:38 - Why Your Git History Is More Valuable Than You Think

00:40:44 - Decision Records: The Architecture Investment That Pays Off

00:43:16 - When Documentation Saved Dennis from a Bad Management Decision

00:44:47 - The Tailwind Layoffs and the Open Source Business Model Crisis

00:46:27 - Guidelines for Consuming Open Source Responsibly

00:49:51 - Why You Should Open Source Your Own Projects

Guest: Dennis Doomen - Microsoft MVP, open source creator (FluentAssertions and more), and coding architect at Aviva Solutions.

#softwaredevelopment #softwarearchitecture #softwareengineering