From Solution Architect to AI Architect

My interest in AI began in 2023 when I first experimented with ChatGPT. Like many of you, I was fascinated by its potential and wanted to understand how prompt engineering could enhance my day-to-day work as a Solution Architect.

That’s how the book Generative AI for Software Architects: How to Use LLMs to Boost Your Productivity was born.

Over time, I began using Copilot to assist with coding tasks. I built an application called ADRGenerator, where Copilot handled most of the frontend (React) development and even supported me with backend (Java/Spring Boot) coding. It was an eye-opening experience as AI was no longer just a concept but a productivity partner.

A significant turning point came when I discovered Claude Code (CLI), an incredibly powerful tool for generating, analyzing, and refactoring code. That discovery deepened my curiosity about how AI could accelerate software development.

My interest grew further as I explored MCP (Model Context Protocol) and Agentic AI concepts through online courses. The idea of intelligent, autonomous agents collaborating to solve complex problems was both fascinating and inspiring.

Next Step

At some point, I realized that to work on real-world, complex AI projects, I needed to go beyond experimentation. I needed a structured learning path —one that builds strong foundations, includes hands-on projects, and ultimately prepares me for AI Architect roles.

So, I turned to ChatGPT for help. I provided it with my background and career goals and asked it to design a tailored self-study plan. After several iterations, I arrived at a refined version of this prompt:

“You are an AI domain expert and career mentor tasked with helping experienced software/solution architects transition into AI architect roles.

The learner has over 20 years of solution architecture experience, strong Java skills, basic Python knowledge, light exposure to AI and strong prompt engineering skills, and wants to avoid heavy mathematical theory while focusing on practical, industry-aligned skills.

Devise a four-month self-study plan ( about 6 hours per week) that prepares them for AI architect roles in Canada/US.

The plan should:

– Recommend specific courses from O’Reilly, Coursera, DeepLearning.AI, YouTube and other platforms covering Python, AI/GenAI, MLOps, LLMs, LangChain, Guardrails, RAG, and cloud AI services.

– Provide weekly structure, practical tips, and certification guidance.

– Present the results in a clear, concise table aligned with the skills expected of AI architects in North America.”

Using that prompt, ChatGPT generated a comprehensive 4-month learning plan with courses, deliverables, and a capstone project.

But I wanted to go one step further.

I gathered over a dozen job descriptions for AI Architect positions in my region and asked ChatGPT to analyze and refine the plan based on the real skills employers are looking for.

I used this follow-up prompt:

“Please review the proposed plan based on the following Architect AI job descriptions in my area.”

Insights from Real Job Descriptions

From that analysis, ChatGPT identified the key skills and technologies that appeared most frequently across AI Architect roles:

What’s Next

I will be refining my 4-month plan along the way, but you can find the version generated by ChatGPT here.

I’ll be sharing my progress, including project milestones, lessons learned, and reflections, and I will keep updating the plan.

This is the beginning of my journey from Solution Architect to AI Architect, and I hope it inspires other architects to take a similar leap and share their experiences.

Stay tuned for my next post!