HEX2017: Energized, surprised and exhausted

MLH is an international student hackathon league, with stops in several countries world-wide. We were thrilled to be the e-commerce partner for the Dutch stop in Eindhoven: HEX2017!

Can we inspire student teams to think on the problems that we are trying to solve? And what can we time-box in 24 hours for a diverse student scrum team?

One of the main topics is the usage of machine learning. We’re experimenting and implementing several algoritms, but one of the key challenges is how to integrate this into our shopping experience?

How can we get a personal connection with machine learning?

As we’re experimenting and exploring on machine learning, the difficulty that everyone is trying to overcome is how to get from a tool to a real connection. Wieden+Kennedy and others are epxerimenting on the field and we feel that this line fits with our view on the developments. The needybot is one of the prime examples of the practical experimenting that is going on now: www.needybot.io

Scrambling on our idea, we’re going throught the needybot kit. We re trying to find the right components, working on the operating system running and figure the configuration. Not an easy task, the libraries of the bot need updating. Also, the operating system is not so easy to run. We get the connection with the needybot eye, the os running and we have a base to start from.

We’re on our way!

SAT, 13:15 hrs We’ve given our pitch to the teams. The other tracks are also interesting and inspiring: VR, using satelite data and …. Who will pick up our challenge? What will their ideas be?

SAT, 14:00–20:00 During the afternoon, several teams approach us to get more insights, pitch their ideas and challenge our assumptions. We’ve identified four wehkamp teams so far… nice catch.

The nice thing to see is that we have four different teams in terms of setup, approach and problem solving. They’re all choosing a different approach, but more than that: they are taking a different approach than we expected. Cool!

We get a lot of interesting questions on our customers, our ideas and give input to their brain storm.

SAT, 20:00–01:00 The teams are prepping for a long night. Using all snacks and dextrose they can find, they all switch from plan mode to execution. Again, every team takes a different approach: one team has made a clear division and all members develop on a part. Other teams have made a different division on labor and have different roles: more design on another team and yet another chooses to pair program.

We leave the teams at their work and make a short round: One team is stuck on the tech details and we try to help without favoring them in regards to the other teams.

SUN, 08:00 The teams have made an all-nighter. Great to see their work and the progress so far. All teams have been able to turn their ideas into a workable base. We see some demos, but every team needs quite some work to get their ideas to a level that they have in mind. A short update and round and we leave the teams to their challenge.

We’re inspired and start working on the arduino’s that we’ve packed. Nice to do on a Sunday morning, but we come nowhere near the teams. But tinkering is a great way to spend your time.

SUN, 13:00 Jury-time! We make our round through the building and visit all teams. To our surprise, there is a fifth team that took up our challenge.

We get five enthousiastic demos. Every team has pushed and prepared a nice demo. All teams show that they’ve put a lot of effort in their ideas and the ideas give us great insights. Tough call to rate their ideas. all have put a lot of effort into it and have come up with great ideas and demos.

One of the teams has come up with several ideas to make needybot more human and more whkamp. They’ve created four different approaches that tie in to a more personalised experience of needy. Based on speech patterns, the bot mirrors words used. It also advises on discounts (even with competitors ;-) ). The team also demos a module that recognises the facial expressions of the people that needybot sees. Great work, the team taken the concept and added some great improvements and worked them out to demoable parts!

The second team took a radically different approach: They created a style platform. People can create a look, share their looks and give other looks points and rating. The demo and presentation are great. they’ve created the platform and the style generators using wehkamp articles. It also contains a style finder based on aI. All in all a cool presentation and great content.

The next team we visit, stayed on the idea of the needybot. They’ve put great effort in tring to get the bot to work on AWS. They’ve also put great effort in facial expressions and demo an implementation that recognises and returns the expression. Nice work in 24 hrs!

We then visit a team that has been working on a different field, but has a nice angle for wehkamp. They gamified the neighbourhood and points and badges are earned. They have an idea to integrate re-utilisation into their game. Nice idea and well worked out.

SUN, 16:00 One of ‘our’ teams get through to the final round. We were not the only judges that were enthousiastic about the teams.

In the end, one of the wehkamp teams wins the finale. Great work and a good idea. But we invited all teams to wehkamp’s warehouse for lunch and thank you for the ideas.