How does In The Pocket tackle the shift to Artificial Intelligence?
During the winter of 2017 nine candidates of In The Pocket, Showpad and MoveUP participated in the first Machine Learning Academy organised by Crunch Analytics. After 10 weeks of training we sat down with Thomas Mons, the Director of Engineering at In The Pocket. We discussed the results and found out what a progressive company like In The Pocket is undertaking in order to surf the ‘machine learning and AI’ wave.
[Interview: Thomas Mons of In The Pocket, 10/01/2018]
Why is Machine Learning and AI important for In The Pocket?
"In The Pocket is a digital product studio. Our goal is to create digital products that make people happy and help companies grow. Within three years all digital products will contain a form of machine learning or artificial intelligence. Because of this, it is extremely important for a digital company like us not to miss the train. Which is why we asked ourselves the question: “How can we start using machine learning and AI?” The first logical step in this process is to make sure we are able to identify the opportunities of machine learning and artificial intelligence in our digital products.
Our initial strategy was to hire the right people who could work out this story for us, but after giving this some thought, we have decided that this was not the ideal plan. Finding talented Data Scientists is not straightforward. These profiles are very scarce, and on top of that they are usually looking for a company with a pure focus on Machine Learning projects. At the same time, Google advised us that it is more efficient to train our own team members so they have a good basic understanding of the basics of machine learning. Ambitious and promising soft engineers are perfect for making this transition. All they need is the right guidance. This prompted us to change our strategy and see what the best training programme would be for our team. "
Why should companies invest in Machine Learning training?
"Our motivation is self-evident: we want to make the best products and this technology allows us to do so. For many people, Machine Learning seems to be something magical, but it is far from impossible to learn the basic principles and models. The main hurdle is often that people are insufficiently aware of the possibilities these techniques offer. Consequently, they are unable to spot opportunities in their projects. Naturally, it is also required that people who do the actual implementation are sufficiently skilled to do so without error.
We are also obliged to embrace this change. Should we ignore the shift to Machine Learning and AI, we will no longer meet the requirements of our users and clients. Customer expectations today extend beyond just a quick loading time of an app or website. The user expects that he will not receive the same offer twice. He wants the right information at the right time, without asking for it.
A final reason is to keep on strengthening our human capital. At In The Pocket we attach a great deal of importance to this value because we depend on the knowledge and drive of our people. We want to continually motivate these strong profiles in their interest in new technologies. That is exactly what makes us strong as a company. "
Why did In The Pocket choose Crunch for the Machine Learning training?
"The match with Crunch was mainly because of the format of their training. Our requirement was that after the training our candidates could immediately start working with basic data science projects. Besides the fact that their training is very hands-on, the flexible format of their training tilted the scale and made us choose for Crunch. Because we could send a select number of people to sessions organised by Crunch, this gave us the chance to try out the training on a smaller scale, and with less stress on our ongoing projects. We decided to send five of our best and finest, which were joined by other candidates from fellow start-ups Showpad and MoveUp.
The third factor was that we were looking for a training that focused on Data Science and would not take up a whole year. At Crunch, our employees followed a half-day lesson per week, spread across 10 weeks. In spite of its condensed nature, we believe that our participants have gained skills that are equivalent to other, much more time-consuming programmes."
What effect did this training have on the participants and the company?
"What is striking is that the interest and drive of the participants have grown exponentially. Moreover, their motivation and confidence definitely got a boost.
Currently, we are experimenting with a project to carry out the learnings of the training ourselves. In the future, we want to launch small features that will make our products even smarter for the user and our clients. Of course, it always depends on the product to determine what is possible with machine learning and AI.
For more complicated projects, we will still call on third parties such as Crunch to assist us. However, it was especially important to gain sufficient knowledge through this training to spot the AI opportunities and to choose the right partners. And Crunch certainly succeeded in that."
Would you recommend the Crunch Machine Learning Training to other companies?
"Certainly! I received a lot of positive reactions from our candidates. The level of the training was exactly what we expected. The pace was there, but that also made it challenging enough for our employees.
A good follow-up for this training would be to give our engineers more knowledge on how they can put these models into production. This being something that of course was outside the scope of the introductory course that they have followed. [CRUNCH: we are happy to announce that we have developed a new module that tackles this issue in depth, the first session is planned to start in the spring of 2018!]
Another benefit is that the formula makes the training very interesting and accessible. It was also one of the decisive reasons why Showpad has participated in this training. Anyway, we will continue to keep a close eye on Crunch and their trainings. "