How I Transitioned from Software Development to AI Engineering?

I became MVP at Microsoft. Today’s newsletter breaks down how I transitioned into AI Engineering.

No AI Roundups today.

In today’s edition:

• Dive Deep Drill— Transitioned from Software Development to AI Engineering

• Build Together— here’s how I can help you

Sponsor Spotlight

Learn how to make AI work for you

AI won’t take your job, but a person using AI might. That’s why 1,000,000+ professionals read The Rundown AI – the free newsletter that keeps you updated on the latest AI news and teaches you how to use it in just 5 minutes a day.

Sign up to start learning.

Dive Deep Drill

If you are a software developer, you already know the whole software building process.

Understand that AI is part of that process.

Without software development, there is no AI development.

How I Transitioned from Software Development to AI Engineering?

As I was a software developer, I understood coding, algorithms, API integrations, front-end and backend, and deployment on the cloud.

I was learning ML and DL along the way, so the transition became smooth.

I made a lot of mistakes, and here is what I learned from them.

I compiled those in 7 steps that you can follow to build AI skills and attract opportunities.

1/ The Mindset [Consistency + Intensity]

Don’t Fall for the AI Hype—Build Real Projects

The hype crowd’s out there chasing shiny AI beta thinking some half-baked tool will rocket their career

Spoiler alert: IT WILL NOT

You’re not here for that noise.

Avoid this and focus on these 2 things for solid work:

• consistency—showing up daily, no excuses

• intensity—focusing on one task at a time

Over time, you will see your knowledge grow exponentially.

Now, you need a learning system to work on daily.

2/ The Learning System

AI mastery takes time, not trends.

I follow this learning system:

— 4 Hours of Deep Work

• No phone

• No talking—just you and the task.

• Coffee or tea’s fine.

• Read, learn concepts and workflows

• Build project/portfolio, Github. [Steal my github profile]

— 1 Hour of Shallow Work

• Share your progress online with the world

• Apply for interviews or reach out to clients

Note: Make a success metric of your learning

You can change the time according to your schedule, don’t put pressure on yourself for sticking for 4 hours every day.

You and I are busy with other things in our lives. Even if you took 10 minutes to learn on a particular day, consider that day a success.

Remember, you need consistency.

3/ Stick to 1 Roadmap

There are plenty of roadmaps on the internet.

They just confuse you and me.

The solution is to focus on one roadmap and not jump on another until it’s done.

You can follow my roadmap - AI Engineer HQ

The best thing is that you can create a duplicate of it and add notes according to your daily work.

Once you complete it, you can now jump around other topics and add it to your learning knowledge base in Notion.

4/ Solve 2 Books [I am recommending]

I learn best through books.

Maybe you learn from videos/blogs or documentation.

However, I recommend/challenge you to solve these 2 books in order:

1. Building LLMs from Scratch by Sebastian Raschka
   https://github.com/rasbt/LLMs-from-scratch

   

2. AI Engineering: Building Applications with Foundation Models
   https://github.com/chiphuyen/aie-book

   

Focus on one at a time. Just don’t juggle around.

5/ Build Projects

This is obvious.

Without hands-on experience:

• No client will want to work with you

• No recruiter will trust your skills

Clients will hire you over anyone else if you have worked on similar problems the client wants to solve.

The same goes with the company.

How do you find good project Ideas?

You Don’t.

That’s right! You have to tweak already existing ideas from research papers and companies who are working on a similar problem.

Here is a simple way I followed earlier.

Later, I will share the new one, but for starters, this is an efficient approach:

1. kaggle’s competition section for problem statements

2. scan featured problems (companies are actually paying to solve these problems)

3. go to the discussion forums of that particular problem

4. skim research papers related to the problem statement

5. frame your problem statement around those research papers

Why this way?

Because hype chasers are drooling over AI trends.

You are grabbing datasets companies actually pay to solve.

6/ Deploy in Production

No AI project is worth it if it is in your local system.

Deployment is needed in any company who are solving a problem using AI.

If you know AI concepts + come up with solutions, + understand the production environment.

You will leave competition behind:

• deploy it live

• let users touch it

• gather use cases if possible

You will be the ideal engineer for any company as you are coming with this much value on the table.

7/ Showcase Your Work

All smart and big companies focus on ROI, which is not Return On Investment.

It is RADIO ON INTERNET.

It means you need to share your work online.

I have been sharing my work online for years now, and one thing that I am sure about is that you will not stop getting new opportunities in your career.

I just became MVP at Microsoft because I was continuously sharing my learning online with the tech community.

Conclusion

Companies don’t hunt for hype riders.

They want engineers with attitude, shipping impactful, profitable projects.

Build one thing—build it right.

That’s it.

If you get stuck somewhere in the process, you can reach out to me or ask in the Discord community.

Want to work together? Here’s How I Can Help You

• AI Engineering & Consulting (B2B) at Dextar—[Request a Brainstorm]

• You are a leader?—join [The Elite]

• AI Training for Enterprise Team—[MasterDexter]

• Get in front of 50k+ AI leaders & professionals—[Sponsor this Newsletter]

I use BeeHiiv to send this newsletter.

Paper Unfold

A series in which you will get a breakdown of complex research papers into easy-to-understand pointers. If you missed the previous ones:

• AI Agents for Computer Use [Paper Unfold]

• Future Of AI - Large Concept Models [Paper Unfold]

• DeepSeek-R1 - A Step Towards AGI [Paper Unfold]

• Agents are Not Enough [Paper Unfold]

• Attention Is All You Need [Paper Unfold]

PS: I am starting the batch for GenerativeAI and AI Agents. Professionals from Amazon, Microsoft, and Accenture are already in the batch. Join The Elite.