My Profile Photo

Wojtek Kulikowski


I am a software engineer writing on ai, startups and product development. I need to yet include words python and cloud for SEO purposes


Winning Pekao Coders

TL;DR. Last weekend we participated in a hackathon organized by a major Polish bank - Pekao SA. The bank provided us and 10 other teams with anonymised data on transactions made in their clients’ terminals. The event was awesome. We won. Part 2 focusing on the development coming soon. You can check out our demo at pekao.kulikowski.tech

Ok, the post went a little story-like but I hope you’ll find it useful anyway.

Couple of weeks ago there was an announcement that Pekao bank will organize a series of hackathons across whole of Poland. The goal would be to make an innovative use of the enormous amount of data that the bank posseses. :chart_with_upwards_trend: It was exciting as I have never come across such a big dataset (even having a year of professional data engineering experience). I immediately signed up.

Fast forward to the event. The contest started on Friday evening when we got introduced to the finance experts and started brewing our ideas. There were two areas in which we could operate;

  • Internal solution for the POS terminals operators (eg. small business owners) :house:
  • External solution for stakeholders that could make a use of the big data insight but lack the terminals that would provide the data (eg. foreign companies that want to enter the Polish market) :office:

We brainstormed a little on Friday with an aim to start the development work on Saturday. The result of our initial discussion was Eventify (original name, right?) - a tool for massive events’ organizers. Think music festivals, fashion shows or car premieres. Given that Louis Vuitton would organize a premiere show in Poland and Warsaw is not the only reasonable choice - where should they do it to maximize their audience engagement? We believe that data about high profile transactions can show us the way, and such insight would be more trusted than a human expertise based on feelings and personal bias.

So this was our first idea, but not the final one. Part of the Saturday agenda was Lean Canvas workshop led by @aelewandowska. We had an opportunity to validate our idea to Lean Canvas methodology and adjust it accordingly. It was some great experience, and we decided to pivot our idea afterwards, due to fact that the problem that we tried to solve was quite artificial and probably would not bring much profit to our stakeholders. We decided to go a little broader and to target the market of small and average businesses (MŚP for short in Polish). At this point we were quite set on analyzing the location data, so we choose to solve the problem of setting up a new business. While doing so it is often tempting to just choose the most popular area or the cheapest one. However, with loads of data gathered by a bank we can analyse where the actual audience is. Louis Vuitton would not open an outlet next to a railway station just because it is popular. They would conduct research to seek a prestigious place with a higher density of their actual buyers. Our solution would automate that process.

When we settled on the idea it was time for the development. We had little over 24 hours to develop a usable implementation of what we created in our heads. For the sake of your sanity and the post’s length I will describe it in detail in a separate post. But it was hard, and we did so many things wrong.

Was fun tho :sweat_smile:

As a result of our (ok, it was my fault) bad coding there was some big crunch on Sunday noon. We had to prepare a presentation and our demo was far from working, server would not respond for some mysterious reasons, and we felt quite miserable. We wanted to have a fully functioning algorithm without cheating but it would just not happen. We started hacking things live on our production server and had no hope whatsoever. Finally, we decided to cut some corners and while the algorithm is technically correct it would show the Warsaw Spire as the best place for your business almost all the time, for any given input. Ooops :man_shrugging: :woman_shrugging:

Relieved that the whole coding wasn’t done for nothing we proceeded to prepare presentation. Now this is tricky. At hackathons as this one, you can have an awesome product but if your fail you 4 minute long (:exclamation:) presentation - your idea loses whole value. We used canva.com (for their minimalistic graphic packs and reliability), recorded a little demo and started rehearsing with one goal - not to blow anything up.

It was just fun. The whole event was brilliantly organised, we learnt a lot and spent 2 days straight learning and creating - a common hobby for each member of our team. During the presentation we managed to make the audience laugh a couple of times and got really good questions from the experts that we have never considered. We were impressed by the app from @wemakebuttons - a competing team that put together a whole game that could be played instantaneously on a smartphone and knew that other teams had some cool solutions as well. We felt that there is a chance for a place on the podium but that was about it.

30 minutes forward, we were screaming in excitement entering the stage as winners :wink:

I hope you enjoyed the story and that you got to learn something from it. The next part for us is competing in the great final of Pekao Coders competition. We look forward to strengthening our architecture and the business plan in the meantime.

If you have any feedback or questions please don’t hesitate and tweet me @wkulikowski1 asap! Part 2 of the post is coming asap too :runner: In the meantime go check out our app (we will turn it off in a couple of days) at pekao.kulikowski.tech and try to score anything other than Warsaw Spire :sweat_smile:

Huge shout out to Pekao SA once again for the amazing event.

Small shout out to my friend Piotr for reading the whole post before publishing :wink:

Till the next one,

Wojtek