Do you need a ph.D. or Math to work with Machine Learning in 2020?

Lima Vallantin
Marketing Data scientist and Master's student interested in everything concerning Data, Text Mining, and Natural Language Processing. Currently speaking Brazilian Portuguese, French, English, and a tiiiiiiiiny bit of German. Want to connect? Envie uma mensagem. Quer saber mais sobre mim? Visite a página "Sobre".


Não se esqueça de compartilhar:

Compartilhar no linkedin
Compartilhar no twitter
Compartilhar no facebook

Não se esqueça de compartilhar:

Compartilhar no linkedin
Compartilhar no twitter
Compartilhar no whatsapp
Compartilhar no facebook

While studying and learning more about Machine Learning and Tensorflow, you will see people saying that you don’t need math or a ph.D. degree to use this library.

In fact, Tensorflow is very democratic, since it’s really easy to use.

In theory, to do Machine Learning and Deep Learning, you need to know Python, have some level of curiosity and a good knowledge of the problem you are trying to solve.

However, some employers request skills that, sometimes, make finding a first experience on Machine Learning very hard.

In order to take a closer look at this situation, I have compiled data from 651 positions with the “tensorflow” keyword published in United States in May/June 2020.

The goal here was to find which skills / technologies are mostly requested.

Please, notice that 651 positions is too little data to be statistically significant. Also, all positions have been extracted from the same website, so bias is huge.

As I mentioned, this is an exploratory project. If you get inspired by this kind of analysis, you can develop a more sophisticated approach to gather more data and to diversify your data sources.

If you want to see how I calculated this, please, use this notebook.

About the data

Data was gathered from Indeed USA division. I extracted from their websites positions from all over the country which matched against the keyword “tensorflow”.

651 positions have been found. I further cleaned the gathered data and extracted common skills technologies from the job description.

The final data is a collection of 651 examples and 35 columns with the skills. When a position requests a skill, this one receives a 1. When the skill is not requested in the position, it receives a 0.

As an anonymization measure, the descriptions, company names and job titles are not disclosed.

Engineering: the most requested skill

When we talk about Tensorflow, we see that 70% of the positions request some experience in Engineering. Without any surprise, python is requested on 69%.

The next most requested skills are:

  • A degree in Computer Science: 58%
  • PyTorch: 47%
  • C: 44%
  • ph.D.: 42%
  • C++: 42%
  • Statistics: 37%
  • Math: 32%

The knowledge of libraries and technologies such as Keras and Spark, and of technologies such as Computer Vision comes after.

R is requested by 22% of the positions.

Natural Language Processing (NLP, my greatest and true love 💖 ) is requested by 22% of the positions.

Do devs need to have a ph.D.?

Now, let’s focus on the positions requesting a developer with Tensorflow knowledge. Only 90 out of the 651 jobs I found request a developer or developers’ abilities.

In the case of these positions, the most requested skills are Python knowledge (85%), Engineering (82%) and Computer Science degree (64%).

Knowing C (58%) and C++ (52%) is also important on this universe, followed by mastering PyTorch (51%) and having a ph.D. (42%).

Statistics (37%) and Math (36%) come just after.

What if I want to work with NLP?

145 positions request some level of Natural Language Processing with Tensorflow knowledge or experience.

If you want to work in NLP field, you will see that the analysed companies still request things like Python (86%), Engineering (84%) and Computer Science (63%).

PyTorch is well known as a friendly library for Deep Learning NLP and, without any surprise, it’s requested by 63% of the companies.

However, Linguistics is only present in 4.82% of the positions! Data Mining and Information Retrieval are also underrepresented with only 15% and 8.96% each.

What do you think?

As I told before, evaluating only 600 jobs all listed on a single platform is far away of being significant for statistical purposes.

But it should helps us to better decide in which technologies or qualifications we have to invest time if we want to work in a particular field.

If you liked this post, let me know what do you think and which technologies and skills you – or your company – consider important when working with Machine Learning.

Não se esqueça de compartilhar:

Compartilhar no linkedin
Compartilhar no twitter
Compartilhar no whatsapp
Compartilhar no facebook

Deixe uma resposta