Why I'm Leaving My Comfort Zone: Staff Engineer AI first Engineer

Source: Dev.to

Background

I’ve been a software engineer for over 18 years, working on distributed systems, data pipelines, streaming infrastructure, and backend services at scale. I’ve built systems that process terabytes of data and architected platforms handling millions of requests. I’m good at what I do.

Why I’m Starting Over in AI/ML

The Company Shift

My company began hiring AI/ML engineers. Suddenly, meetings were filled with discussions about embeddings, fine‑tuning, and Retrieval‑Augmented Generation (RAG) systems. I could nod along, but I had no idea what they were actually building. After 18 years of experience, I couldn’t contribute to the most important projects happening at my company, and that hurt.

The Claude Code Experience

I started using Claude Code and it was a game changer: proofs‑of‑concept that used to take days were done in hours, new features were delivered 2–3× faster, and going from 0 to 1 on projects became almost effortless. The problem? I didn’t understand how it worked. I was using AI without building it, unable to explain its decisions or judge whether solutions were genuinely good or merely convincing.

When someone asked, “How does Claude Code actually work?” I had no answer. That moment made me realize I needed to return to fundamentals.

The Realization

I’ve always believed that understanding fundamentals lets you solve complex problems. When I truly learned distributed systems—beyond just using Kafka, I understood partitioning, replication, and consensus—I stopped merely using tools and started building systems. That shift made me valuable.

I need to do the same with AI: not just use it, but understand it and build it.

Goals

• Become an AI‑first engineer – be part of the conversation, not just a nodding participant.
• Build AI systems from scratch – understand models, architectures, and trade‑offs.
• Architect production‑grade AI solutions – design, deploy, and maintain robust AI services.

Specific Objectives

1. Learn how models like Claude actually work.
2. Build AI systems from the ground up.
3. Lead or contribute to AI projects in production.

Plan

I’m committing to learning AI/ML fundamentals from the ground up—no shortcuts, no “just‑use‑the‑framework‑without‑understanding” approach. I’ll document the journey publicly, sharing wins, struggles, confusion, and breakthroughs.

• Week 1: Begin with core concepts (linear algebra, probability, optimization).
• Ongoing: Dive into model architectures (transformers, diffusion models), training pipelines, and deployment strategies.

I have no fixed timeline; the commitment is to the journey itself.

Call to Action

If you’re also making a career transition into AI/ML, I’d love to hear about your experience in the comments. Let’s learn together.