Why I Am Not Trying to Be an AI Researcher Anymore

July 14, 2026

For a while, I thought I wanted to go deeper into AI research.

It sounded right. AI was popular, everyone was talking about it, and it felt like the field where “smart people” were supposed to go. Very healthy reason to choose a direction, obviously.

But the more I stayed around it, the more uncomfortable I became.

Part of it was imposter syndrome. I kept comparing myself with people who had better math background, better research experience, better papers, better everything. I know comparison is not always fair, but knowing that does not magically stop the brain from being annoying.

Another part was fear of achievement. That sounds strange, but I think it is real. In AI, I felt like I had to constantly prove that I was impressive. Better results, better models, better benchmarks, better keywords. It started to feel less like curiosity and more like a performance.

I do not hate AI. I still respect it. I still use it. I still think machine learning is useful and interesting. But I do not think I want my main identity to be “AI researcher,” at least not now.

I feel more grounded when I build systems.

Backend services. APIs. Databases. Testing. Software that has clear behavior and real workflows. Things that break in understandable ways, then force me to debug them like a normal suffering developer.

I also still care about security, privacy, and applied cryptography. Those areas feel closer to the kind of thinking I enjoy: systems, protocols, constraints, edge cases, and actual engineering trade-offs.

Maybe one day I will return to AI from another angle. Maybe through privacy-preserving systems, federated learning, or secure software. But I do not want to chase a field just because I am scared of looking less impressive.

That is not a good way to build a career.

So for now, I am choosing software engineering, backend systems, testing, and security-minded applications.

Less glamorous maybe.

But it feels more honest.