How I went from Software Engineer to AI Engineer in less than 6 Months

Where It Started

I started working at JP Morgan full-time as a Software Engineer in 2023, right after graduating college. After a summer and a semester as an intern at the Palo Alto location, I knew this was the best fit for me.

I still remember the day I got a call to join the firm. I was so excited to finally be back in the Bay Area. I was in something called the SEP Program, which is a two-year rotational program for new grads.

I was lucky, and got the opportunity to choose the team I wanted to join. That was when I heard of my team - AIOPS. It was a team building a revolutionary AI Internal chatbot for JP Morgan employees to use.

This is where my journey began.

The Moment AI Clicked

There is no 'aha' moment. It is a gradual, stressful, challenging and ambiguous process. You are learning things and finding solutions to problems not many people have. You are reliant on completing and executing work yourself, with no documentation on it. With a team of 6, you can't exactly ask for help and drain the team's bandwidth. You have to figure it out on your own, and that is what makes it so hard, but also so rewarding.

I self taught myself a lot. I spent weekends studying the entire project and understanding parts not meant for people without a PhD. I read research papers, watched YouTube videos, and read blogs. I also had to learn how to use the tools and platforms we were using, which was a whole new world in itself. It was a lot of trial and error, and a lot of late nights.

But at the end, I slowly realized our standups became less and less confusing.

Eventually I proposed solutions, sped up embedding generation, helped chunk and upload data to our vector database, and was the head of the data quality on our team.

I handled new clients, onboarded them, and handled the relationship they had with our team. I handled their data, helped make sure their data was ingested and also helped run the weekly scraper. Eventually I also became the one who ran and maintained the airflow pipelines to help check the data quality.

Not only was I proposing new solutions for clients and facilitating the progress of those tickets, but I was also making our product efficient and scalable.

This role and responsibility was a huge jump from my previous software engineering role, and I had to learn a lot on the fly. But it was also the most rewarding experience of my career so far. I felt like I was making a real impact, and I was learning so much every day.

The Learning Curve

It felt like I was thrown into the middle of the ocean and expected to learn to swim to shore. Nothing really made sense. Sure, I did AI projects at school and had a pretty high-level understanding of ML, but my focus as a software engineer didn't require use of AI yet. I prioritized LeetCode over projects and research in college, and here I am in a new world, completely lost.

To be an AI Engineer, you have to be comfortable with ambiguity, and you have to be comfortable with not knowing things. You have to be comfortable with learning on the fly, and you have to be comfortable with failing. You have to be comfortable with not having a clear path, and you have to be comfortable with not having a clear solution. You have to be comfortable with not having a clear answer, and you have to be comfortable with not having a clear direction.

This is what makes it extremely rewarding. Driving new developments without help, mastering making POCs on the fly to explain ideas, acting like a PM to showcase product enhancements, and having an ownership mindset will set you apart from the rest.

It is easier now than ever to be an AI Engineer. Claude, Gemini, and other LLMs have made it so much easier to learn and understand AI. So if this is something you want to do, you absolutely have no excuse. You learn by doing, and if you aren't working at an AI startup, you can always do a project yourself!

Advice for Engineers Considering the Switch

Whether or not you like AI, it is mandatory to be up to date on the latest technology. While it may not replace all tech jobs, it will certainly be integrated into them.

As an engineer, you are expected to keep learning and adapting to new technologies and AI is no different. We already have co-pilot everywhere — debugging and writing code for us.

If you are considering a switch, make use of your company's resources and learning initiative! Those are always there to help you. In addition, I believe in a 'learning by doing' mindset. So, if you have the time, try to use the AI tools available and create something cool! There are hundreds of sample projects to learn about LLMs and Agents on GitHub, and you can always find something that interests you.

AI opened the world to so many possibilities, and I am excited to see where it takes us. It is a challenging but rewarding journey, and I am grateful for the opportunity to be a part of it. Anyone can learn and use AI to positively impact the world, and it is something we should all be excited about. The future is bright, and I am excited to see what we can accomplish with AI.

More details in my portfolio! If you have any questions about the role or the journey, feel free to reach out to me on LinkedIn or email!