Data Scientists Must Revisit Their Toolsets: Let Me Explain | by Pedram Ataee, PhD | Jun, 2022

Whether you are a data scientist or want to be one, you must revisit your toolset

Are you a data scientist looking for a job in non-software enterprises? Do you need to extract insight from a large dataset in a short amount of time? Do you want to evaluate whether your idea is solvable by artificial intelligence? If your answer to any of the above questions is “Yes”, this article may help you.

I recently got interested in the new wave happening in the AI world named “No-Code AI”. In the AI community, we still do not have a consensus on how to precisely define the no-code AI technology, though. Products such as DataRobot define the no-code AI as an Iron Man Suit for data scientists. However, other products such as Sway AI define it as an Elder Wand (Harry Potter’s wand) for domain experts and data scientists.

IMHO, the latter definition is closer to the main objectives of no-code AI technology: (1) expediting the process, (2) reducing the costs, and (3) improving the explainability. The no-code AI technology has been born to help enterprises adopt AI technology. So, if you want to work in large non-software enterprises or extract insight from a large dataset in a short amount of time, you must become familiar with the world of no-code AI.

Let’s explain more.

If you want to work in large non-software enterprises, you must have no-code AI platforms in your toolset.

If you have read my previous articles on AI strategy and data strategy, you already know that many challenges exist for an AI project to become successful especially when it runs in a non-software enterprise. For instance, enterprises need to run many experiments to identify problems that can be efficiently solved by AI before hiring a large data science team. Hiring a large data science team is costly and it is risky for an enterprise to build a data science team with a myriad of unknowns. The no-code AI technology is the answer to many of those unknowns. As a data scientist, you can create value for non-software enterprises using no-code AI platforms without relying on other experts such as data engineers who were needed in other projects.

As a data scientist, you may say that you still prefer coding; however, enterprises expect results as fast as possible. If you want to develop everything from scratch or code within a Jupyter Notebook, you may not be able to meet the deadlines or, also, explain things well enough to the executive team. The no-code AI platforms help you, here.

There is no doubt that no-code AI technology is rising. Be prepared to use it efficiently whether you are a data scientist or a domain expert.

If you look at the no-code AI landscapes, you will find out that there are different types of no-code AI platforms: (1) Industry-focused such as Accern, (2) Technology-focused such as Lobe, (3) Data science tools such as H2O, and (4) Category Z such as Sway AI. I will explain these categories, especially Category Z, below.

The industry-focused companies are those who choose an industry or sector (e.g., financial services) and aim to provide all the required AI-based tools in that sector. The technology-focused companies are those that select a specific AI technology (e.g., computer vision or natural language processing) and aim to build their company around that. The third group is those companies that build tools for data scientists mostly focused on AutoML.

The last group or Category Z is those companies that aim to build tools for data scientists and domain experts. While they incorporate AutoML in their product, they create insightful dashboards for their use cases. It is hard to build a product that is being used by everyone but that is what will revolutionize the market. BTW, Category Z is the name that I call them. Z refers to something that comes in the last and finishes the story.

So, please choose your no-code AI platform wisely otherwise you will again be trapped in the unnecessary complexity.

I am very excited about the opportunities that no-code AI platforms bring to this field. As a whole, even with the advent of no-code AI, there will always be a need for data scientists and coding. While no-code AI technology will resolve many challenges, I still believe a fully-fledged AI product needs coding and wouldn’t happen in a matter of days or weeks. So, enjoy using a new tool in your toolset but please don’t misuse it! 😊

If you like this post and want to support me…

Whether you are a data scientist or want to be one, you must revisit your toolset

Are you a data scientist looking for a job in non-software enterprises? Do you need to extract insight from a large dataset in a short amount of time? Do you want to evaluate whether your idea is solvable by artificial intelligence? If your answer to any of the above questions is “Yes”, this article may help you.

I recently got interested in the new wave happening in the AI world named “No-Code AI”. In the AI community, we still do not have a consensus on how to precisely define the no-code AI technology, though. Products such as DataRobot define the no-code AI as an Iron Man Suit for data scientists. However, other products such as Sway AI define it as an Elder Wand (Harry Potter’s wand) for domain experts and data scientists.

IMHO, the latter definition is closer to the main objectives of no-code AI technology: (1) expediting the process, (2) reducing the costs, and (3) improving the explainability. The no-code AI technology has been born to help enterprises adopt AI technology. So, if you want to work in large non-software enterprises or extract insight from a large dataset in a short amount of time, you must become familiar with the world of no-code AI.

Let’s explain more.

If you want to work in large non-software enterprises, you must have no-code AI platforms in your toolset.

If you have read my previous articles on AI strategy and data strategy, you already know that many challenges exist for an AI project to become successful especially when it runs in a non-software enterprise. For instance, enterprises need to run many experiments to identify problems that can be efficiently solved by AI before hiring a large data science team. Hiring a large data science team is costly and it is risky for an enterprise to build a data science team with a myriad of unknowns. The no-code AI technology is the answer to many of those unknowns. As a data scientist, you can create value for non-software enterprises using no-code AI platforms without relying on other experts such as data engineers who were needed in other projects.

As a data scientist, you may say that you still prefer coding; however, enterprises expect results as fast as possible. If you want to develop everything from scratch or code within a Jupyter Notebook, you may not be able to meet the deadlines or, also, explain things well enough to the executive team. The no-code AI platforms help you, here.

There is no doubt that no-code AI technology is rising. Be prepared to use it efficiently whether you are a data scientist or a domain expert.

If you look at the no-code AI landscapes, you will find out that there are different types of no-code AI platforms: (1) Industry-focused such as Accern, (2) Technology-focused such as Lobe, (3) Data science tools such as H2O, and (4) Category Z such as Sway AI. I will explain these categories, especially Category Z, below.

The industry-focused companies are those who choose an industry or sector (e.g., financial services) and aim to provide all the required AI-based tools in that sector. The technology-focused companies are those that select a specific AI technology (e.g., computer vision or natural language processing) and aim to build their company around that. The third group is those companies that build tools for data scientists mostly focused on AutoML.

The last group or Category Z is those companies that aim to build tools for data scientists and domain experts. While they incorporate AutoML in their product, they create insightful dashboards for their use cases. It is hard to build a product that is being used by everyone but that is what will revolutionize the market. BTW, Category Z is the name that I call them. Z refers to something that comes in the last and finishes the story.

So, please choose your no-code AI platform wisely otherwise you will again be trapped in the unnecessary complexity.

I am very excited about the opportunities that no-code AI platforms bring to this field. As a whole, even with the advent of no-code AI, there will always be a need for data scientists and coding. While no-code AI technology will resolve many challenges, I still believe a fully-fledged AI product needs coding and wouldn’t happen in a matter of days or weeks. So, enjoy using a new tool in your toolset but please don’t misuse it! 😊

If you like this post and want to support me…