Typing DNA - My First AI tool in Software Engineering

The first project where I worked with an AI tool was an online exam proctoring system, and the tool we used was something I hadn't seen before: TypingDNA.

The idea behind TypingDNA is clever. It recognizes a person's unique typing pattern, essentially the rhythm and cadence of how you tap on a keyboard, and uses that as a biometric identifier. Think of it as a two-factor authentication layer, but instead of a code sent to your phone, it's the way you type.

That time was relatively new to me in Software Engineering. I had only about 6 months of experience. I conducted R&D on Typing DNA for the team and built a proof of concept. I still remember that my manager was very happy with my work and the R&D output.

I don't exactly remember the entire implementation process, but it was pretty easy to implement. They had well-documented APIs.

In our project, we integrated it into the student registration flow. When a student scheduled their exam session, we captured their typing pattern at that point. Then, when they actually sat down to take the exam, the system verified that the same person was behind the keyboard. It was a lightweight but surprisingly effective way to ensure exam integrity without invasive camera setups or complex proctoring software.

It was a genuinely fun tool to work with, and for its use case, it worked really well. I'll link the tool below. I'm not entirely sure how relevant it is today, given how much the AI landscape has evolved, but back then it felt like a neat piece of technology that solved a real problem in a non-obvious way.