Ever wanted to know more about the people behind your favorite Enterprise AI platform? You're in luck — every few weeks, meet some of the humans at Dataiku working every day to ensure customers and users find success on their path to Enterprise AI. This week, we spoke with Jean-François Yuen and Pierre Bailly-Ferry who are both software engineers working on our engineering team. We asked them to introduce each other, in writing and in video, so make sure to watch them play our fun “who is the most likely to...” game to discover who cheats when playing board games, writes cleaner code, and much more.
Can you please describe each other’s role within Dataiku’s engineering team?
Jean-François: Pierre was part of the Genki Dama team (yeah, we have some weird team names) when he joined Dataiku, where he did a great job promoting best practices and performing unit testing. He is now managing his own team — known as Designer — responsible for all the things related to the flow in the software, I am sure he’ll do an even better job!
Pierre: Jean-François is part of the HAL team, our R&D team specialized in dealing with everything related to machine learning. For example, he developed the ML Diagnostics, a small notification that helps our users understand why a model is not as efficient as it should be.
What makes you like working with each other?
Jean-François: Pierre is very knowledgeable about everything on Spark, Java, and general best practices. He is always willing to help when asked (or not) and gives really good advice and feedback. Moreover, Pierre likes doing a clean and good job, and this shows out in the software.
Pierre: Jean-François is always helping others. If you have a question, even unrelated to his work, he will always find the time to help you.
Can you give us an example of a project where you particularly enjoyed collaborating?
Pierre: In my first month at Dataiku, I worked on a particular bug that needed validation, the kind of bug with more buzzwords than real words on its title (like proxy, GCP, and more super technical terms that require specific competencies). I had a hard time setting up the test environment due to its complexity. I told him my frustration during the coffee break, and he decided to help me. We did some pair programming and, less than half an hour later, the cluster was up and ready. So we took a second coffee.
Jean-François: We worked together on the Model Document Generator feature, which was quite complicated because of all the external dependencies. Pierre made sure that every part of the feature had a unit and integration test. When expanding the features, it’s easy to build off of what has been done.
Is there a topic that can easily lead to hours of debates between the two of you or a topic you like to argue about together?
Pierre: Cooking. I like to spend hours preparing my own food with (mostly) quality ingredients. Jean-François could eat anything, as long as he does not have to cook anything.
Jean-François: I think talking about software quality and good practices is a never-ending discussion between us, but we agree on most points. We both have a desire to improve our product — not just on the features, but also on tooling and process for developers. The real difficulty is applying the theory to the real-world code in Dataiku across multiple distributed teams. Lunch is also a major topic between us. I usually eat quite late and love having a happy meal with a few nuggets, but that is not to Pierre’s taste unfortunately. It’s hard, almost impossible, to make him change his mind on that.
According to you, which Dataiku value (or aspect of the culture) does Jean-François for Pierre and Pierre for Jean-François represent the most?
Pierre: “Get things done.” If anything goes wrong, you can be sure that Jean-François is already fixing it.
Jean-François: I think it’s more about doing a good job. Pierre really emphasizes software quality and will ensure that what is shipped meets his high standards. By that, I think it reflects on the current Dataiku engineering culture: It’s all about making the software better for the users first.
You are two experienced engineers who have worked at other companies. Is there something that you believe is unique about being an engineer at Dataiku?
Jean-François: Dataiku is a company that has grown quite fast in the last few years and is still expanding quickly. At first, the priority was on shipping new features (you’ve got to make a product!), but now it’s also about polishing and making life easier for users. So we talk a lot about user experience and quality. In that sense, Dataiku will not hesitate to delay a feature or a release if it doesn’t meet our expectations, and that’s quite unique, because it’s not about delivering at all costs and rushing the development, but being happy about a job well done.
Pierre: What I love about Dataiku is that you get to be in charge of your features from specification to delivery. You have to begin by understanding the client's needs. Next, you have to define the functional scope with a product manager. Then, with your team, you are in charge of providing the technical solution. Finally, we gather feedback through users' test sessions to ensure that the feature matches the client’s expectations. This is also when you get the best ideas for future improvements.