Whether you are a student gearing up for a Career in Data Science/Machine Learning or a person transitioning into this field, you need to follow some general guidelines and they are as follows:-
1) Brush up Probability, Statistics, Calculus and Linear Algebra:
Get familiar with these topics as they make the foundation for understanding the algorithms. You should know what you trying to do, so that when you get suck, you can brainstorm out of the problem.
Here is my favourite book for brushing up important concepts.
Link: https://mml-book.github.io/book/mml-book.pdf
There are some concepts regarding Reinforcement learning such as Bellman Equation, Dynamic Programming etc. which are not mentioned in the book given above. However, the following book covers them in great detail.
Link:
https://drive.google.com/file/d/1uqBEvPE9nr03r7fNp5vdiQJcahvTo3Uz/view
2) Choose your language: Python or R.
Both are equally good. If you are choosing both good great but it is quite challenging for a newbie. Start with one, become proficient in it, then pick another.
3) Practice makes a man perfect:
For example: music. Pick a singer you like say, Shawn Mendes. Take all his songs and find out which note he sings most often. Gather common datasets and work on them like Boston Housing dataset or Chicago Crime dataset. Later on, practice on Kaggle and participate in hackathons.
4) Get a good working knowledge of Data Structures:
These are core Computer Science subjects and familiarity with these is a must. Competitive Coding is important, so get your hands dirty on Leetcode and Hackerrank.
5) Pick a domain or two to specialize:
Here comes the main part, your specialization. It depends on one’s interest. Computer vision, NLP, Speech, Finance are some of them.
6) Be proficient with cloud platforms:
Cloud Technology is one of the hot skills these days. Every cloud service given below provides specialization courses and certification exams. Finish the courses, learn the skills, ace the exams and finally earn a certification.
- Amazon Web Services
- Google Cloud Platform
- Microsoft Azure
- IBM cloud
- Nvidia CUDA
7)Make a GitHub account:
There is a video given below explaining the basics of GitHub. It is a version control platform but it can do a whole lot more.
Link: https://www.linkedin.com/feed/update/urn:li:activity:6554650688364875776
8) Do some unique projects:
That’s your portfolio. Projects reflect your expertise and experience level. Also, create a portfolio website using GitHub pages. It’s free and does an excellent job.
9) Keep your LinkedIn profile up to date and stay active.
Networking is very important in today’s world. Your LinkedIn profile is the first thing a recruit sees and the second is your Resume, so design and draft them well. Internships and Research assistantships will do wonders for you.
10)Business Acumen:
Get comfortable speaking to Stakeholders, Business owners and the people at the top. Understanding the problem from the top down and then working out how the data fits into the overall “business” is something I’m hearing is absolutely the key.
11) Machine Learning and Deep Learning theory:
The following link contains examples of theoretical questions generally asked in an interview.
Link: https://medium.com/subhrajit-roy/cracking-the-machine-learning-interview-1d8c5bb752d8
12)Learn Tableau:
Strong visualizations are also important. There are some technicalities which are always difficult to explain to people with non-technical backgrounds. Tools like Tableau play a huge part in solving this problem too.
13)Stay updated:
Try to develop a habit of reading blogs and posts on Medium, Hackernews, AI Time Journal etc. You can also read research papers at ResearchGate or Google Scholar. Be a part of a Machine Learning or Data Science Community to get the information on the latest technological trends.
Now just keep in mind that you guys don’t have to let this stuff overwhelm you. It’s a lot and I know that. Make at least a one-year target to do all of this stuff and focus on quality, don’t rush through things.
All the Best!
Anything to add, please?
