The Oppy is reconnecting with some prominent NYU Stern alumni with our ‘Graduate’ series. Each month, we’ll interview notable alumni that have made an impact on their community and industry.
This month, The Oppy is featuring Romain Fouache (‘17), COO & CRO of Dataiku. Romain and I sat down in Paris and New York, respectively, to talk about his career as a data scientist and business leader. Read on to learn more about Romain’s love of math and his perspective on the future of “everyday AI,” including one hot topic that may interest you: how to best prepare yourself for careers in data and AI.
This interview has been edited for length and clarity.
Romain, you’re the first graduate we’ve featured who attended TRIUM, the Global Executive MBA program jointly issued by NYU Stern, LSE and HEC Paris. It sounds fascinating. Can you share more about the experience?
Yes, it’s a collection of some of the best professors from each university and a pretty exciting package. Over 18 months, you get to study in five or six different cities with the same cohort, from 10 days to two weeks at a time. The Executive MBA provided both “globetrotter” and cultural diversity aspects that might be a bit more difficult to get within a single school. It was an amazing experience.
How much engagement did you have with NYU throughout your program?
We spent just two weeks in New York but then, depending on which cities we were going to after, we had professors visiting from NYU. But the time in New York City itself was the same amount of time that we spent in the other cities we visited like Shanghai, Chennai, and London – just a couple of weeks.
Switching gears to your current role at Dataiku, COO & CRO are two sizable positions in themselves! What’s it like managing the responsibilities of both roles?
It might seem a bit schizophrenic from the outside, but I think that both are actually very aligned. COO and CRO are two of the key components to allow and ensure the scale of the organization. CRO is the scaling, making sure that we’ve got adoption in the market and that we can transform the lives of our customers. COO is always making sure that everything internally is aligned and in place to allow the growth of the organization.
When I joined Dataiku, just under four years ago, we were a hundred employees. Now, we’re close to 1,000. The internal change is just as major as the external change. We changed scale, like moving from one neuron to a brain – it’s something entirely different in itself. Supporting organizational growth from both a CRO and COO standpoint allows me to have a more comprehensive view on the challenges that we’re facing and how we can analyze them holistically. It makes no sense to scale up our presence in the market if we cannot scale internally, and it makes no sense to scale internally if we cannot scale our presence in the market.
That makes complete sense. I took a look at your LinkedIn profile and saw that you started your career as a data scientist after receiving a master’s in applied mathematics, and then you pivoted to technology startups. So I have a couple of questions related to that.
Firstly, what was it like working as a data scientist back in 2002?!
Well, as a start the names have changed. “Big Data” was not a thing, “AI” was a bad word, and Data Scientists were called Data Miners. 30 years prior, you might have gone to a logistician or logistics statistician, and it would have been basically the same job. You have large amounts of data and you need to make some sense out of it so that there can be a business outcome. The underlying principles are the same and, actually, the underlying principles of data science are hundreds/thousands of years old, and the technological principles date back to the 1950’s.
You could say the role itself hasn’t changed but for two things. The first one is the pervasivity of the technology. In the early 2000s, the friction of being able to access the technological capability to do the work was huge. 15-30 years ago, only a handful of very large corporations could do it and you needed to be very specialized to access it. Today, everybody can. You go online, you have a credit card and you can start using it. There is no cost of entry to be able to perform data science. That’s big change number one.
Big change number two is the pervasivity of the skills. Ultimately, anyone can do it, it’s not just about being a math PhD. It’s about understanding the underlying principle and then applying it. It’s the same as when, in the 80s, people thought that computers were for a handful of ivory tower computer scientists. Today, it’s the same for data science. Even you mentioned before the interview started that, at NYU, you got hands on or close enough experience with doing data science, even though you’re not from a technical background. So it’s not a question of the specific domain or specific skill set, it’s something that can be applied to every domain, in which every skill set can benefit. It’s not math, it’s business.
Yes, because if you can’t actually translate data science into something meaningful from a business perspective then it’s not useful (… shout out to my Data Science for Managers Prof. Claudia Perlich!)
What prompted your shift from a practicing data scientist to business leadership roles in technology startups?
It was my loss of naivete on math. I majored in math, and I felt when I got out of my university days that math was the solution. If you had a problem with lots of data, you’d have a way to process the data that would give you the best outcome. I started working on these problems, and by actually being in the field and talking with companies I changed my mind. One of our main customers at the time was a major outdoor advertising company. We were working on pricing optimization, which I initially thought was a purely objective mathematical problem until I realized that, no, it was a business impact problem. I realized that math is part of the toolbox that you have but, ultimately, it does not de facto lead to a single objective outcome. Math can give you answers, but will not frame the question. What’s the business context? What’s the outcome you’re looking for? Who are the people that are involved and how can you help them actually do better and be better? Instead of just saying, ‘Ok, it’s math – it’s easy, it’s objective.’
And on top of that, even when you know the question, there is no one way to ask it to the data. There are a million of them, which means there’s no objectivity because you can get an objective answer but you will always have a subjective question. If you don’t understand the business part of it, you will never understand the question that you are actually asking. This is when I realized that, yeah, math is great – and I still love math as much as I did at the time – but it’s only a very small part of the picture.
If you actually want to be able to leverage all of the amazing technologies that are in front of us, you need to understand the bigger picture – and its business. It’s business in the sense of how organizations deliver the results they need to deliver, regardless of whether it’s sales or marketing or human resources. So I grew in my subsequent role to work on these types of challenges, always with an element of bias towards being data driven and leveraging math to achieve results.
Do you think your background as a data scientist gave you and the companies where you worked an edge?
I think it might have, because it was a time when data didn’t have such a prominent role. When I graduated from college, Google was not a public company and Amazon was just a bookseller. So this pervasive culture of data as we know it today did not exist and, as a consequence, it was not as much of a focus. In the early days, it was growing a bit, but was seen as the realm of IT. So I had an edge because I took an approach that was not mainstream yet.
Now people understand the value and the importance of data, and, as we said earlier, you’re in business school and you’re learning about data. At Dataiku we have many hugely talented data scientists that come from business school. Because, ultimately, the edge is not about deep math anymore. Like using a computer, your edge does not have to be about being able to code in assembly, it’s about being able to leverage the tool and understanding the best of what can be done with it. That’s the skill.
So you spoke a little bit about the changes in data and AI over the years. The underlying fundamentals are the same, but now the access to and the importance of data science is ubiquitous. Let’s take that a bit further: where do you think careers in data and AI are going in the future?
Yes, that’s a good question. I think it’s going to go in two directions. First, having specific people dedicated to internal analytics and data is something that is a long-lasting trend. We’re going to have more and more people doing that, but I think it’s going to be an ever decreasing proportion of the total number of people actually leveraging data and AI. I think using data and AI is going to become a daily activity by almost everyone, in the same way that using a computer is.
So a company can be moving from, let’s say 100 data scientists to 50 data scientists, but more importantly they’re going to move from 10 people not being data scientists but using data and AI to 10,000 people doing it. This is where I think the trend is going. It’s more about allowing everyone to become exceptional with AI by using AI every day, then just the core skill of the data scientist that will always be important, always be essential, and always be in demand.
So by daily use, for example, do you mean like how Dataiku’s platform has algorithms built in and tools that would help someone non-technical carry out certain tasks?
The easiest comparison is with Excel. Most people we know use Excel on a daily basis. They weren’t raised as computer scientists, but they found something that supported what was needed to do the task at hand. Whether it was to have an address book, or to better forecast their cash flow, or to track inventory, I think what we’re looking at is natural movement like that. There are many mundane tasks that can be automated very simply by using this type of technology as soon as, and this goes back to our previous discussion, someone knows the question to ask. It cannot be done outside in a lab because you need to know what your question is, but as soon as you know what your question is, the rest of the things are easy.
We work with companies where, when they wanted to do a simple analysis, they needed to go to IT, ask for an extract, and wait two weeks for it. Then, they had to ask someone with a specific skill set to look at it and copy and paste it into five different spreadsheets and one macro developed by an intern in 1990. Only then could you gain any insights from it. All of these friction points are the typical things you could do today with the tools and with almost no additional skills in one hour. And for many companies it takes one month. There is friction all down the line that you can streamline by democratizing some of the tools and some of the technologies.
Another analogy, very simplified, is to think about how we used to send letters 50 years ago. You were a business person and you wanted to send a letter. So, you would dictate your letter and someone would type it and send it to the internal dispatching office that would send it to the post to be received at the recipient’s dispatching office before it was delivered. Today, you open your computer, you type an email, and it goes. So what used to take two weeks and require six people actually takes two minutes and requires one, and I think that’s the evolution we are going to see with data and AI.
Of course, if you want to send a big package you’re gonna have to go through the full shipping process. AI is going to be the same. You still need the data scientists to do the big stuff, and that makes a whole lot of difference, but the just daily things, everyone will be able to do by themselves.
Those examples illustrate really well where it’s going, thanks.
As you might expect, as soon-to-be graduates of Stern, we are focused on our next job opportunity. I’d love to hear your pitch about Dataiku and why some of our students should consider it as a potential employer.
Yes, of course. What we discussed previously is not only my core belief, but it’s the belief we’re pursuing at Dataiku. My perception is that we are part of the trend of the next big change in the way companies actually operate. There is not a single decision that cannot be augmented with the right data, and companies are now coming to the realization that, not only, it’s actually possible to achieve business results at scale with AI and that it’s not just for Silicon Valley big tech companies and, second, it has to and will become a standard part of the operating system.
This is the shift Dataiku is enabling organizations on. We are working with people and organizations to help them build better rockets, just some who just want to better forecast cash flow for next week, covering everything from the mundane to the moonshots. It’s changing the way they operate and I don’t think any company is going to be able to continue to thrive without having been able to empower their people to make everyone extraordinary. So, if you want to be a part of it, a part of that story, then Dataiku is the place to be.
But it’s also not just about what we do, there’s how we do it. I’ve never enjoyed a role as much as I do at Dataiku. I think our people are nothing short of extraordinary themselves. We are one of the best rated companies on Glassdoor, and Dataiku is very true to itself. It’s however not necessarily for everyone: Dataiku typically values collective success above individual achievement, which may not be a fit for all. So, if what you want is to be with a team of amazingly talented people, make a difference and rejoice in the collective success then Dataiku is an amazing place to work. We’re hiring a lot, in all positions: data science, talent acquisition, customer success, sales – if you’ve got talent we’ve got the role.
… I’m sold! Jokes aside, I love that you touched on the cultural aspect because that plays a big part in keeping employees engaged every day.
Now we’re going to have some fun. I’m going to take you through my “rapid fire” questions:
What is one word that describes you: Versatile
Favorite professor at Stern? It would be difficult to say anyone other than Foster Provost. He literally wrote the book on Data Science for Business.
What are you reading right now? La Promesse de l’Aube by Romain Gary, a French author from the 20th century
Finish this sentence: At 7am you can find me… in my bed
What is a cause you’re passionate about? I’m passionate about helping people excel and doing the best of what they can do
What do you do for fun? Spend time with my kids
Favorite movie, TV show or podcast? Favorite movie: The Prestige by Christopher Nolan
What’s one thing you can’t do that you want to learn? Speedreading!
If you could meet anyone alive or dead, who would it be? Let’s go back to my math background. I’d love to meet Kurt Gödel. He was, in my opinion, the person who wrote the most important work of the 20th century, proving that there is no such thing as universal mathematical truth and actually demonstrating it, not just philosophically.
That’s really cool, did he win a Nobel Prize?
No, he didn’t (there’s no Nobel Prize for maths!), but he was friends with all of the people who were at the heart of the invention of computers. To show how much of a logistician he was, he died of hunger for fear of being poisoned.
Oh my God.
The only person he trusted to cook his meals was his wife. She got sick and was hospitalized, and Gödel died of malnutrition, because he just couldn’t bring himself to eat something else for fear of being poisoned – that’s how driven by logic he was.
Oh, wow, what a brilliant man and what a tragedy.
Time to get back to business here. How do you think your time at business school influenced your career today?
I think the greatest benefit I received from the business school is the diversity of cultures, of origins, and of experiences that I met. The thing is, usually as we go through our career, our environment becomes narrower and narrower. We meet people who are more and more and more like us and, as a consequence, we lose a sense of understanding of who we are, because we can only compare ourselves to people who are in a very close dimension to where we are. And so, you just measure yourself in one dimension. But, actually, when you engage with diversity, you realize that there are 20 other dimensions that are there and that you can be seen, and you can assess yourself in many different ways.
The last question I have for you is: what advice would you give to fellow MBAs who are interested in either tech startups or careers in data and AI?
They can send me their resume, to begin with!
My advice is to understand the right level of balance when it comes to data and AI. If you are doing an MBA and you’re not already a data expert, then it’s not about becoming such an expert. It is, however, about understanding the underlying business principles that actually drive whatever organization you are going to be in next and thinking about how they can be augmented with data. So I would say, as for any other topic, stay open, stay aware, educate yourself – don’t try to be an expert but do be someone who actually understands the basic definition behind it and you will always have a job. There will be quite a lot of them in the future.That’s great advice, thanks so much for sharing your story with us, Romain. Sternies, you can find Romain on LinkedIn and learn more about Dataiku at dataiku.com.