Getting into AI, a decade ago

Today marks a decade since my first day of BS and start of my journey into AI. Seeing students having 1st day of university or joining after lock down, this week made me more nostalgic.

On 16th Sept 2010, I joined COMSATS Lahore, with lot of hopes and dreams of learning and excelling at Artificial Intelligence and Robotics. I spent 11 hrs at campus that day, spending a lot of time in the awesome library, reading books on those topics.

This was way before it was so popular and mainstream like today. Before AlexNet and the boom of Deep Learning. Before Big Data was "big". Before Coursera. Years before TensorFlow, PyTorch, GANs and AlphaGo etc. Before Harvard Business Review called Data Scientist, "the sexiest job of the 21st century".

Back when the field was a bit abstract and mostly focused on symbolic and "good old-fashioned AI", search & planning, expert systems and agent based systems, with Machine Learning just beginning to see widespread application.

When Microsoft was developing Kinect using Random Forest for Pose Estimation. When Apple was developing Siri with Signal Processing Transforms based speech-recognition, and was about to amaze people. When Google was struggling with improving statistical machine translation based Google Translate. When I heard people saying stuff like: "I'm waiting for Weka to support Support Vector Machines".

I'm glad that I still got opportunities to work in the domain throughout the decade. I started programming intelligent systems during first semesters and mailed the AI teacher requesting to let me attend AI course in 2nd year instead of final year, which he gladly accepted. I got some good teachers who involved me in projects, research and competitions related to AI and Robotics. People used to tell me that I'm wasting time and AI and Robotics have no future, specially in Pakistan but I was determined. I got a very relevant job here after graduation and later did my MSCS from Georgia Tech specializing in AI, during which the industry matured a lot here and I got a very relevant job again. I got to utilize a lot that I've learnt.

I feel that I could've done so much more with all that passion, energy and time, if I were to start my AI journey as freshman in modern day with all this excessive amount of courses and learning resources, papers, codes, developments in field, well defined topics and tech stack etc.

The little positive side was developing a good foundation, programming expertise and knowledge of old but still useful techniques. It also felt nice and special to be doing it when it wasn't so popular and mainstream. Now it just doesn't feel as cool and competition maybe intimidating for beginners.

A takeaway from all this is that, if you are a student interested in AI and Data Science then you should feel lucky to be able to get deep into it at early age in this era and take maximum advantage of the opportunities.