In my last article, I boldly declared taking on the world while working on a mysterious project. Well, unfortunately for me, the world has won that battle.
My focus in recent weeks, or even months at this point, has been on the domain of keystroke dynamics and its commercial application in authentication and authorisation. I have to say, I find the field absolutely fascinating and was initially surprised by how easily we can be recognised just by how we type. Aside from the interest in the domain, it was also a playground to develop machine learning skills while solving a real-world problem (after careful consideration, I decided the world doesn’t need yet another cat image classifier).
The journey has been thoroughly gratifying: constant learning and developing new habits that helped me to stay on track. Re-discovering the power of writing. Reading through scientific articles published on the topic of keystroke dynamics, siamese networks, and triplet loss. Working with data. Running systemised experiments. Appreciating the power of an LLM as a sidekick. Working with a programming language that was not my strong suit before (python). Going from zero to a commercially-viable model. Learning about how to approach starting your own venture. And the most painful lesson of them all: always double-check if someone else does not have a patent on what you’re trying to develop; the sooner the better.
The journey does not end here though. Nor do I regret embarking on it. Even if this alley turned out to be a dead-end, it was still worth taking for all that I have gained along the way.
At the same time, I’d lie if I said this discovery was not demoralising at first. But I got over it. As will you if your project doesn’t pan out.
Where to next, then?
Maybe the notebooks filled with notes from the project will spawn a flurry of technical articles, perhaps even a course. Or maybe not.
Maybe I’ll start another, unrelated venture. Or maybe not.
Maybe I’ll go back to being employed because I do miss coding on more regular basis. Or maybe not.
Leave a Reply