From Electrical Engineering to AI SaaS: My Career Pivot Story

How I transitioned from a BEng in Electrical Engineering at KUET to building production AI SaaS products in London — the skills that transferred, the gaps I had to fill, and what I'd do differently.

The Unexpected Foundation

When I studied Electrical & Electronic Engineering at KUET in Bangladesh, I didn't expect it to be the foundation of an AI career. But looking back, the overlap is significant:

• Signal processing → directly maps to time-series ML
• Control systems → the feedback loops in RL agents
• Circuit analysis → debugging complex system interactions
• Mathematics — linear algebra, calculus, differential equations → the backbone of ML

The hard skills transferred. The mindset transferred even more.

The Gap: Software Engineering

What EE didn't teach me was production software engineering. I knew how to write code that worked on my machine. I didn't know how to:

• Structure a codebase for a team
• Write tests
• Deploy reliably
• Think about scalability
• Handle authentication, payments, user data

I spent the first year at UEL filling these gaps aggressively.

What Accelerated My Learning

1. Building real products, not tutorials

I stopped following tutorials after completing the basics. The moment I started ContentForge AI — with a real goal, real users, and real bugs — my learning compounded.

2. Reading other people's production code

Open source repositories (FastAPI, LangChain, Next.js examples) taught me patterns that no tutorial covers.

3. Virtual experience programmes

J.P. Morgan and Walmart's virtual programmes gave me exposure to enterprise engineering practices and something concrete to put on a CV.

4. Embracing failure

ContentForge AI had 47 production bugs in the first month. Each one taught me something a course never would.

The Skills That Matter Most (In Order)

1. Communication — explaining complex AI systems to non-technical stakeholders
2. Problem decomposition — breaking hard problems into solvable pieces
3. Speed of learning — the specific tech stack matters less than how fast you learn new ones
4. Shipping — a mediocre product that exists beats a perfect product in planning
5. Technical depth — eventually you need it, but breadth gets you started

What I'd Do Differently

• Start building products in year 1, not year 2
• Contribute to open source earlier — it builds reputation and skills simultaneously
• Document everything — this blog is 2 years late

If you're considering a similar pivot: the path from engineering to AI is shorter than you think, and the foundations you already have are more valuable than you realise.