From CRM Developer to AI Engineer: My Career Path

Where It Started

My journey into AI engineering started in a place most people would not expect: Salesforce Marketing Cloud. I spent years building email campaigns, CRM integrations, and marketing automation workflows at Dyson. It was deeply technical work, but it was not what anyone would call AI.

Looking back, though, that background gave me skills that turned out to be directly transferable to AI engineering in ways I did not appreciate at the time.

The CRM Foundation

Working with enterprise CRM platforms taught me things that serve me every day as an AI engineer:

• Data modelling: CRM systems are fundamentally about structuring and querying complex data. This translates directly to designing schemas for AI applications
• API integrations: I was connecting dozens of systems via APIs long before I connected my first LLM. REST, webhooks, authentication flows, rate limiting: all the same patterns apply
• Automation at scale: Running email campaigns for millions of recipients taught me to think about queuing, batching, error handling, and monitoring
• Business logic: Enterprise platforms require understanding complex business rules and translating them into code. AI applications need the same skill

The Turning Point

The moment everything changed was in late 2023 when I started experimenting with the OpenAI API. I had been following AI developments casually, but it was not until I actually built something that I realized how much my existing skills applied. My first project was automating a content categorization task that had been eating up hours of manual work. It took me a weekend to build, and it worked better than I expected.

That weekend project turned into a second project, then a third. Within a few months, I was spending every evening and weekend building AI tools. Within a year, AI engineering had become the primary focus of my work.

Skills That Transferred Directly

If you are considering a similar transition, here are the skills from my CRM background that transferred with zero modification:

• Python: I was already writing Python for data processing and automation scripts
• SQL and databases: Working with Supabase and PostgreSQL feels natural after years of CRM database work
• JSON and API design: Every AI application is essentially an API that wraps model calls
• Testing and QA: The discipline of testing email campaigns before sending them to millions translates perfectly to testing AI pipelines
• Stakeholder communication: Explaining complex technical concepts to non-technical teams is the same whether you are talking about CRM workflows or AI pipelines

Skills I Had to Learn

Not everything transferred. Here are the areas where I needed to build new expertise:

• Prompt engineering: This was entirely new and took months of practice to get right
• Understanding model capabilities and limitations: Knowing what AI can and cannot do reliably requires hands-on experience
• Vector databases and embeddings: Concepts like semantic search and RAG were completely new to me
• Cost optimization: LLM API costs can spiral quickly. Learning to optimize token usage and choose the right model for each task was essential
• Evaluation and scoring: How do you measure whether an AI system is performing well? This is a distinct skill

How I Made the Transition

I did not make a dramatic career leap. The transition was gradual and intentional:

• Started with personal projects to build skills and a portfolio
• Found opportunities at Dyson to apply AI to existing workflows
• Built increasingly complex systems, from simple API wrappers to multi-agent pipelines
• Documented everything I built and learned (including this blog)
• Connected with the AI engineering community online

The best career transitions do not require starting from scratch. They build on what you already know while adding new capabilities that unlock the next level.

Advice for Others Making This Transition

If you are a developer working in CRM, marketing technology, or any enterprise platform and considering AI engineering, here is my advice:

• Start building immediately. Reading about AI is not the same as building with it
• Pick a real problem you understand deeply and solve it with AI
• Do not wait for permission. Build projects on your own time if needed
• Your enterprise experience is an asset, not a limitation. Many AI engineers lack production experience with real-world data and business constraints
• Focus on building end-to-end systems, not just model prompting. The engineering around the model is where the real value lives

The AI engineering field is young enough that there is no single "right" path into it. If you can build reliable software systems and you understand how to work with AI models, you are already qualified.