Modern Day Data Science
Perception vs. Reality on the “21st century sexiest job.”
Last week I talked to a senior software engineering manager from a well-recognized company. While speaking to him, he said that Data Science would be automated soon. I was curious and asked him why he thought it would be automated. He said that there are many easy-to-use tools that can build, train and tune machine learning and deep learning models automatically. I agreed with him but also said that building and tuning models are just a tiny part of the Data Scientist's job.
Automation is ubiquitous not just in Data Science but in any domain where machines can perform repeatable tasks. Currently, there are many open source and proprietary libraries like H20.AI, Azure AutoML, AWS SageMaker, etc., which automate training and tuning models. But, the expectation of what a successful Data Scientist brings to the table is much more than just training and tuning Machine Learning models.
A modern-day Data Scientist can extract the voice of people or customers from their data in a million different ways and use the information to drive the business value for their company.
After attending multiple conferences, meeting different Data Scientists, and working for many companies, I want to list the top non-technical skills needed for a successful Data Scientist that cannot be automated in this modern world.
A natural and most common trait for a successful Data Scientist is the curiosity for data. I would say Data Scientist is a Sherlock Holmes in the world of data. Curiosity leads them to meet many people and learn different perspectives on solving problems. For example, I once had lunch with Adam Cheyer, co-founder of Siri, at the A.I. ReWork conference in San Fransisco. His take on Conversational AI completely changed my perspective while solving Conversational AI problems.
Passionate Data Scientists tell people what cool things they are trying out — people get really excited about that. They’re not going to say you tried and you suck, they’re going to say, “Wow, you actually did something. That’s cool!” — Hilary Mason
2. Patience and Persistence
Good Data Scientists are persistent in many ways. It is not necessary that millions of people find it hard to spot patterns and Data Scientists just take minutes or hours to find patterns and solve the complete problem. It is good to be true, but it rarely works out that way. Real-world data is messy, and it is sometimes frustrating to understand and clean it. Others may decide to move to the next problem, but persistent Data Scientists will stick to solving the problem.
Data Scientist cannot be a 9–5 employee. Some part of the problem is always in their mind!
3. Purpose and Pride:
It is good for data scientists to come up with multiple ideas and experiment with the latest technologies. But, the problem comes if that is the only motivation. A Data Scientist should not run complex models when a simple solution is good enough. The objective must be to solve the problem and make a business impact.
Data Scientists take pride in “how much business impact they made and how many people they helped than just how many advanced techniques they used.”
New technologies will come and go, which is why people now believe that whatever shines in the tech world won’t remain for a longer time. I also agree that programming knowledge, cloud technology, database skills, etc are essential and might get automated because of technological advancements. However, the automation of these skills will only accelerate but won’t automate the whole process of solving problems for Data Scientists. Moreover, as some parts of Data Science will see automated processes but the field of Data Science and the demand for Data Scientists will continue to grow.
All you need is passion, patience, persistence, purpose, and pride!