It was a really good experience to conduct an interview of Ali Mustufa Shaikh with Harsh Navani!
Ali is a humble 23 year old, with the tagline ‘I build Human Networks’ (not Computer Networks).
He has done massive work for the tech community and is the youngest Intel Software Innovator in India, a Microsoft Student Partner, a Google Certified Educator, a Co-organiser of GDG Cloud Mumbai, the Organiser of TensorFlow UserGroup Mumbai, the lead of the APSIT IEEE branch and an Advisory Member at IEEE Collabratec among other things. He has also been a Global Influencer at Google Crowdsource for two years.
We talked about his journey right from the beginning till date, took some advice for beginners looking to start with AI and some other personal questions. The following questions are some of the most important ones that we asked Ali during the interview. Enjoy!
How can one get started with Artificial Intelligence (AI)?
Ali suggested a ’top to bottom’ approach:
- Don’t learn AI at first, learn software development. Learn how web sites, APIs and servers work. This will help in applying AI, because at the end, one needs to create business value.
- For AI, start with non-coding concepts & go through Machine Learning for All on Coursera to get some non-technical and business perspective on AI.
- Use tools like Teachable Machine, AutoML or go thorough Google AI experiments.
- Now start learning frameworks like PyTorch or TensorFlow.
- Start working with smaller data sets and move on to Big Data. (One might want to learn about Cloud technologies while working with Big Data.)
- Once one has applied models in applications and has some experience, then get into Math, optimization of functions and deep dive into concepts.
- Now one can make their own models.
This advice coincides with Jason Mayes’ advice on how to get started with AI. I was fortunate enough to interview him as well and I’ve documented that in an article as well.
How important is Math in AI and how can one learn it?
Math is the core of AI and Ali added that he always tells people that the ‘M’ in ‘ML’ (Machine Learning) stands for ‘Math’.
He gave different perspectives on how one can go about it:
This is for someone who is not interested in working with softwares, but is more interested in researching about softwares.
For such a person, Math is very important. One cannot start researching in AI without it. Learn Math & AI related concepts. Following a ‘bottom to top’ approach is the way to go.
For such a person, in the beginning, Math is not very important and going deep into an AI technology is also not necessary. One needs to learn tools that will enable them to apply AI technologies in their projects.
As one keeps using these technologies more and more, the need for optimizations or customizations increases. Here is where Math is required and should be delved into deeper.
So basically a ’top to bottom’ approach is the way to go.
Which resources can one refer to for AI?
Some amazing people/organisations that Ali mentioned:
What is ‘Machine Learning Fast Track’ (MLFT)?
MLFT is a course that Ali created when he was visiting different colleges and saw that there was a big gap between what students wanted to learn and what was being taught to them.
He teaches the course (with certificate), but you can do it yourself as well (without certificate) as the course is open source and can be found on his GitHub profile!
Project based learning versus course based learning
A combination of both is the way to go, according to Ali!
Only learning through projects will not help one gain general knowledge of a topic which might be of real use or might be required in emergency situations.
Only learning through courses is not helpful as one will not have the practical experience of building something on their own.
Ali also recommended reading programming books. He does that himself. Even though technology is advancing at a rapid pace, the core concepts usually remain the same, so it makes sense to read books.
How to maintain a balance between project based learning and theory?
- Start with theory
- Pick up a dataset from Kaggle (if it is an AI project)
- Try to apply the theory (might not work out)
- Learn how to implement it (theory or tutorials)
- Implement it
- Learn to optimize it
- Implement optimizations
- Add more features by repeating above steps
Ali also suggested having a 80-20 blend, with 80% focus on the project and 20% focus on the underlying aspects of the project. This way one will not learn only what is required in the project, but will also have some knowledge about the concept.
What value do course and professional certificates provide?
He made his opinion on certificates very clear through two posts (#1 and #2) on LinkedIn. Nothing more needs to be said!
How should one handle mistakes, especially at higher positions?
One should acknowledge the mistake, reflect upon it, apologize to people and try their best to not repeat it again. They should not make promises, but try their best to not make the same mistakes again and again.
Ignore the haters. One’s followers do understand the fact that they made a mistake and if they acknowledge and apologize for their mistake, people do forgive them.
How can one manage their time?
Ali is very intentional with where he puts in his time and he avoids ‘rescue missions’.
‘Rescue missions’ are situations or things which have been delayed so much that all available time has to be devoted to that one task, leaving all other things aside. He really avoids those patterns from building in his schedule.
From a hiring perspective, which qualities in a person are the most impressive?
- Humbleness & politeness
- Ability to adapt, learn and implement things
- Ready to learn and implement fast (ie, a good attitude & mindset)
- Some technical knowledge (will be covered up later through mentoring, if lacking)
These were some of the most important questions that Ali answered. I hope you enjoyed reading and learnt something! The entire interview is available on YouTube.
You can follow Ali on LinkedIn, Instagram, YouTube and Twitter.