During my time at AWS, I worked as a cloud support engineer supporting various AI and ML services. My experience with these tools came primarily from troubleshooting, configuration support, and general guidance. I helped customers solve problems, but I never fully understood how they were using these services or what was happening behind the scenes.

That changed with the rise of generative AI and Amazon Bedrock. Suddenly, I had the chance to explore not just the tools I knew, but new ways to design and manage AI systems.

What followed was an intensive certification journey: four AWS certifications in five months. Going through the certifications made me see cloud, AI, and security in a new way, and the insights I gained are useful for any organisation leveraging AWS.

Partner Requirements and Preparation  

The partner tier required me to complete four certifications quickly. I followed a structured sequence, starting with the basics and moving to more advanced AI topics. The key was not just using study resources but reviewing answers carefully to understand why each choice was correct or incorrect, while also developing a deeper understanding of the core domain concepts behind each certification. This approach helped me think more clearly about problem solving and system design.

Learning Through Client Work 

At the same time, I was working on client projects. Real deadlines meant I could not just memorise information – I had to apply it immediately. Learning about systems in theory is very different from putting them into practice. Working on live projects taught me lessons that studying alone could never, showing me the consequences of decisions in real situations.

Exposing the Gaps  

This combination of study and client work revealed gaps in my knowledge. Years of support experience taught me how to fix problems but designing systems from scratch required a different way of thinking. Setting up pipelines, managing databases across locations, and handling migrations transformed exam questions into real decisions with real impact.

I assumed my background would carry me. I knew the AI services, but generative AI changed how things work. The focus is no longer just about calling pre-built models; it is about guiding AI to give useful, accurate results. Success depends on framing the right questions and directing the system responsibly.

Discipline Over Usage  

The most important lesson was that AI is not just about using tools but about applying them thoughtfully. Anyone can click a button. The difference lies in setting up systems to achieve goals, checking outputs, keeping data safe, and controlling access. Understanding infrastructure and cost became critical, especially for AI systems where usage can grow quickly without careful management.

What Certifications Revealed 

Each certification highlighted things that support experience alone could not: the architectural thinking, the cost consciousness, and the operational rigour that separates someone who uses AWS services from someone who builds sustainable systems with them.

Security by Design  

Security decisions now come from understanding the why behind controls, not just the how. Before certification study, I would have used simple static credentials because that is what I had seen done. Now I understand how secure authentication reduces risk and can explain it to clients. Multi-account setups now focus on limiting potential impact rather than creating unnecessary complexity. These lessons came from thinking about security at a deeper, architectural level.

FinOps as Strategy 

The FinOps certification shifted my perspective, it completely changed how I view cost. It is no longer about small optimisations but about helping teams and organisations manage costs strategically. This includes planning accounts, tracking usage, and choosing flexible system designs that adapt as needs grow. Reducing onboarding time for clients from weeks to under an hour is an example of applying this thinking to real work.

A New Way of Thinking  

The shift in thinking is simple but significant. Before certifications, I asked: “Will this work?” Now I ask: “Will this work, stay secure, scale, and make financial sense for the client?” The foundation of how I approach problems has permanently changed.

Process Over Badge 

Certifications validate knowledge, but preparation transforms thinking. I noticed this shift during a client conversation about database replication: instead of reaching for documentation, I found myself reasoning through trade-offs I’d internalised from practice exams. You can’t cram your way to that. The repetition of reviewing incorrect and correct answers, understanding why right answers were right, builds mental models that persist long after the exam. The growth is in the process, not the badge.

Familiarity Isn’t Fluency  

Experience teaches familiarity; structured learning teaches fluency. Five years at AWS taught me what tools do. Certifications taught me when and why to use them and how they connect. Support work answers, “How do I fix this?” Certification thinking asks, “How should this be built?”

Connected Architectural Thinking 

These certifications also gave me frameworks to mentor others, not just share personal stories. They add credibility when recommending system designs, and ensure security, cost, and scalability are considered from the start. Every solution is now evaluated against best practices. Scattered knowledge became connected thinking, shaping everything I build for clients.