The Beginning

I really don’t know what title I can put on myself. Although, in my current comapny, I am considered to be the Lead AI Scientist but to be honest, I don’t know what that means.

My background is machine learning and deep learning. I have worked on multiple aread of machine learning (some which I don’t have a paper). In my master’s thesis, I was using machine learning to differentiate between a special type of evoked potentials in the brain called the “Error Potential”. At that time, I was obssessed with orthogonality and especially when it came into a non-orthogonal space.

When I entered my PhD in 2013, it was the year when the AlexNet came out and therefore, Deep Learning was the craze. There was no pytorch, but there was torch7. There was no tensorflow, but there was caffe. I needed to up my game and eventually, I got my PhD with torch7. Tensorflow was open-sourced just one year plus before I finished my PhD.

Now, lonmg story short: I am a person that likes to think meta and design software. When I entered the production ML world, I didn’t know about design patterns, agile, kanban, XP, software architecture and many other concepts a person must know to do professional softare development. Comin to know about all these things, it hit my mind that what do I want to do for my career? Where do I see myself in a few years time.

After some research and surfing the web, I came across an interesting paper “Machine Learning Architecture and Design Patterns”. This paper was a real eye-opener for me and I decided that maybe, this is the direction I want to go towards: becoming an Machine Learning Architect.

Now as an architect, one of the first things is to understand is that what does an software architect do? and how do we translate the practices of software architecture design to machine learning problems.

I’m hoping to shed more light on this.