top of page

Why PhDs make good data science professionals

Updated: Mar 9, 2022

Data science professionals are currently highly sought-after profiles in a large number of companies. Working conditions and salaries are generally very competitive, and the jobs are interesting for those who want to bring meaning to data. Although the sector is still quite young, employers are already looking for candidates with a higher and higher level of seniority. In addition to technical and technological skills, they are also looking for flexible profiles with excellent adaptability, with the aim not only of grasping the rapid changes in the sector but also of deriving interesting business opportunities from them. We had the opportunity to speak with Yotta Academy founder and CEO, Sacha Samama about the skills in demand and the expectations towards candidates. We intend to determine their adequacy with the doctoral skills, which are real assets in these professions.

Which opportunities for PhDs in data?

For several years now, business and tech pundits have been describing data as the new black gold, whether right or wrong. Data Scientist's job was crowned Job of the Year for four consecutive years by the job search and recruitment site Glassdoor in the United States. It remains in the top 10 today. In France, hundreds of job offers appear on LinkedIn for positions such as Data Scientist, Data Analyst, Data Architect and Machine Learning.

"In terms of sectors where demand for data professionals is increasingly strong, one can think of industry, for instance in supply chain management, especially in retail," says Sacha Samama. "There is also a strong interest in these profiles in the various fields of healthcare. We are already starting to see an emergence of specializations at the level of company functions or even sectors of activity." Beyond the buzzwords, data professions offer very concrete career opportunities.

In order to keep (or enhance) their competitive edge, companies large and small are trying to make the shift and become data-driven, i.e. guided by data in their processes, their business decisions, but also in their services or products. As data professions continue to define themselves and mature, Chief Data Officers are taking their place at the top of organizations' hierarchical structures, alongside traditional C-level executives such as Chief Operating Officers (COOs) and Chief Financial Officers (CFOs).

This transition, however, presents significant challenges. Some of them are addressed in the 2019 edition of the "Big-Data and AI Executive Survey", which is conducted by professor and frequent Harvard Business Review contributor Thomas Davenport and Randy Bean, CEO of NewVantage Partner. The survey gathers the responses of 64 data and IT executives from major global corporations (Citigroup, Sanofi, Ford Motors, Credit Suisse, etc.). It shows that only 31% of respondents describe their organization as data-driven. This is a significant drop from the 37.1% reported in the 2017 edition, despite ever-increasing investments in transformation initiatives. Still, only 5% of respondents cite technology issues as a challenge to business transformation, while 40.3% point to a lack of coordination and organizational agility, and 23.6% point to cultural resistance within teams. This suggests that the solution is more likely to be found in the medium to long term; and that there is no easy, ready-made solution.

As Davenport and Bean put it, "Thinking about data as an asset is a new phenomenon for most companies. Traditional businesses were not data-driven in the sense that emerging digital competitors are". It is nevertheless an interesting context because it places data and AI team members at the interface of technological and human issues.

Do companies see an advantage in calling on PhDs in order to unleash the potential of their data and AI initiatives? "Yes, and more and more. "Says Sasha Samama. " Whereas more junior profiles have been favoured in the past, we're now at a point where employers need more seniority in their ranks, especially from a technical or business point of view. [...] Given the choice between a junior profile and a doctorate, a company will find that there are several reasons to favour a PhD, including their analytical abilities, in-depth thinking and, in some cases, experience in handling data during their thesis". In this context, it remains to be seen whether they will be able to make use of doctoral skills beyond technical or research expertise.

Which skills to work in data science?

Data science jobs require a good knowledge of statistics and mathematics. It is often not necessary to be a mathematician, but a more than correct mastery remains fundamental, for instance, for selecting the appropriate models and algorithms according to the context (CNN, k-NN, SVM, decision trees...) Excellent programming skills are also essential. Among the most in-demand scripting languages are C++/C, Python, and R. However, a host of other languages are also sought after, and it is important to take note of specific expectations when applying. A fine knowledge of good development practices (CI-CD, DevSecOps, agility,...) is also essential.

AWith the industrialization of data also came another type of profile, the machine learning engineer, a profession that the Yotta Academy offers to teach in a bootcamp-style school. "This calls for a whole host of notions around object-oriented programming, advanced Python, code packaging, technologies specific to continuous integration, continuous deployment, cloud computing, etc. The industry is changing at an accelerated pace, and technologies can quickly become obsolete (less than 2 years). A good training today must, therefore, be agile and quickly stick to the state of the art. "adds Sacha Samama.

Although technical skills are considered essential, soft skills prove to be indispensable as well, as the field constant adaptation from workers. Among the most commonly requested skills are analysis, communication, teamwork, and method. So where do PhDs stand in regards to these?

The core competencies of PhD holders

Obviously, PhDs are experts. They cultivate their expertise over the years, for example by working on their thesis, writing articles, participating in conferences, organizing symposiums or doing a postdoctorate. Nevertheless, they have more than their expertise to offer in the work environments where they work. Over the last few years, Adoc Talent Management has conducted various studies on doctoral skills. Among them, CAREER, conducted with 4500 participants in France, and PhDetectives, with more than 1200 respondents in Canada. Both studies provided original data on issues such as the career paths of doctoral graduates and their satisfaction with their doctoral program. They also provided an overall picture of the skills developed during doctoral studies and demonstrated the existence of a pool of core, or central, competencies. Core competencies are common to all PhDs, regardless of their research discipline, seniority, or any other factor.

Here are the competencies that were most frequently mentioned in the respondents' responses in Canada, and which are therefore most likely to be found in the PhD population: research method, analytical skills, scientific and technical expertise, in-depth thinking, independence, time management, and perseverance.

Among the skills most frequently cited in the CAREER study in France were: scientific expertise, communication, project management, scientific watch, autonomy, perseverance, adaptability, and the ability to work in a team. These are all vital skills for data teams in charge of enabling their company to become "data-driven" and establishing a "data" culture internally. "They are usually very comfortable when it comes to data reskill. They have a good learning capacity and retention, also from a technical skills point of view. Even a doctor who is not a mathematician can learn the job," says Sacha Samama.

Zoom sur Sacha Samama

After initial training in mathematics, with a focus on data science in his final year, Sacha worked as a data scientist at Quantmetry, a French firm specializing in artificial intelligence. Convinced of the added value of data science professionals in companies, but also of the need for a more agile and practical training offer to meet the needs of the job market, Sacha founded the Yotta Academy. The school offers the first Machine Learning Engineer training in France in bootcamp format. The Yotta Academy is preparing a "tech summer school" for PhDs and PhD candidates, with the aim of improving their employability in the sector.

Have you ever thought about a career in AI or as a machine learning engineer specifically? How would you describe the training currently available for these jobs? Feel free to share your thoughts!



Los comentarios se han desactivado.
bottom of page