VK Hackathon 2019 (as it was)

Last weekend (September 27-29), “VK Hackathon” was held, at which my team and I were lucky to attend as participants. In this post, I would like to talk about how it turned out and how hackathons are cool.

Firstly, about our team

We have a fire team and I would like to thank everyone, I did not expect that everything will turn out so powerfully.



The team itself ( Already in St. Petersburg ):

• I wrote a backend and recommendation model
• Arina - designer everything you can
• Seryozha - frontendil like a god
• May (in fact, he is Imom) - collected information, showed his communication skills and kept in good shape

July 25th

It all started on this day, it was then that I created our konfu, in which we discussed the tracks and selected features. Initially, our choice fell on the track "Travel", a case from "Aviasales". We thought through everything in detail, understood what features we wanted to fasten and have already seen our product.

August 23-25

Nevertheless, we gave birth to our application and registered our team. And then a languid expectation began, because the results of whether or not they should have taken us should have been published only on September 9th.

8 September

One of my teammates (Seryozha) calls me at 14: xx and says that we have problems. The organizers wrote to us and informed that we did not have enough points to become participants in the Hackathon on the selected track. The fact is that there were a lot of registered teams on the "Travel" track. We were offered to change the track to “Fintech” or take a case from “TASS”. We decided that the TASS case does not suit us at all and our soul does not lie with it, so we began to understand the track "Fintech". I took a look at the Mastercard case, while Serezha liked the PSB case more. But in the end, we chose a case from PSB. We wrote a decision and filed the application in 2 hours and managed to send it 20 minutes before the deadline. Everything was even more ironic given the fact that at that time I was riding a bicycle in the suburbs, and Arina and May could not fully help us.





(Here I went along such a suburb and came up with a new case.)

9th of September

All day was a continuous expectation. Last year, the results of the selection were posted at 10 a.m., and this year only at 19 p.m.). But the results did not please us at all. We were neither in the track "Travel" nor in the track "Fintech". Well, in sorrow we decided that next year we will be more lucky. And our connection with the Hackathon was supposed to end.

10 September

In the evening I was looking for some kind of group, I only remember that it was somehow connected with the case championship, where my friend Sasha called me and scrolling through the entire list of his groups (as many as 31 pieces), for some reason I saw two hackathon icons. One of them was known to me for a long time, but the new one in the title had “... ON BOARD”. It turned out that this was a group to which the hackathon participants were added, for organizational issues. But I did not understand how this could happen, because our team was not on the list. I checked the list again and it was still empty there. After which I wrote to the organizers all the way, and by some miracle, they just forgot to add us to the list. As a result, we became participants in this wonderful event)

September 11-26

All the following days before the hackathon, we thought who, what and how would do it. We decided that we would not depart from the application and would follow it. However, in the application, we indicated that we will implement the application on the VK Mini Apps platform and the whole problem was that in order to write an application for this platform, you need to write in React (JavaScript framework). At the time of writing the application, we thought that somehow our designer Arina would cope with this, but it turned out to be much worse - she had no experience with him. In the end, the hard fate of studying React fell on Seryozha (I am truly sorry for him), and it was complicated by the fact that he had to learn it in a week. We decided that the best way to learn is the project, so we wrote an application that will contain similar elements, but naturally we did not implement our application from the hackathon, because it is forbidden!

September 27

Well, that long-awaited day has come. We arrived at the Manege by 14:30, as was said by the organizers, registered, received a merch and went to our table. They laid out the laptops, found some water, walked, took a picture, and were hipsters (drank smoothies). And at 16:30 the opening began. Of course, our hands itched already to start writing code, but we were presented with all the cases and at about 17:45 we were released to write the code. We had 42 hours to write the application. I understood what needs to be implemented from the backend and by late evening (about midnight) I finished writing the basic methods, May found sites for parsing, Serezha collected the application’s skeleton, and Arina painted the necessary icons. Part of our team (Arina and Seryozha) was already thinking of going to sleep, but then the curator approached us, we exchanged a couple of phrases and he went on. Something slipped through May’s head and he went to talk to him. They talked for a long time, about 20 minutes, I already thought that May delved into the banking business. Then we already talked with the curator as a team and led us to change our idea, this did not affect the application mechanics, but the idea now clearly clung to and had an understandable business model. After that, Arina and Serezha went to bed, and May and I stayed for the night.

September 28th

All this night I was mainly engaged in collecting data from sites. In the morning, by 6 o’clock, May fell asleep under the video about the presentation, and I struggled with sites alone. Then we had breakfast and May changed with Arina and Seryozha, going to sleep. Something did not cut me to sleep and I tried to help Seryozha. Since I didn’t rummage over React, for the most part I acted as a duck, which explains the code and thus finds errors. By lunch, we had almost nothing in the app. And then I was a little depressed. I decided to add a backend, spent some time on it, I also decided to write a recommendation model, surprisingly, but it even worked with indicators (0.6 / 1.0). Although the indicators were not very good, but the examples were acceptable. Then I see how things are going with our frontend, and our application begins to take on a human look. Then I felt a rise. By the end of this day, something similar to a viable application was already looming. But the next day is over.

September 29th

I didn’t feel the lack of time, I understood that in principle we were in time, the flight was normal. We relocated from our table to very comfortable sofas, where we spent the whole night. Hours until 6 in the morning, we slowly debugged the last bugs and prepared a presentation. By 8 o’clock I was already definitely squinted (by this time I hadn’t slept for 46 hours), but I fixed bugs, albeit slowly. I had to chop off the model, because she broke another functional, and it was very difficult to fix it with stuck eyes. Then I decided to lie down at least for an hour, because I had already debted the code, presenting it as a picture of the screen in my mind. Well, I didn’t sleep for long, a little less than an hour. Then we started solving issues with the presentation. At this point, the application already had a pleasant appearance, which I was very happy about. As a result, we uploaded the presentation and link to git 8 minutes before code freez (the time when you can’t change anything already). I was pleased with the result and was in a combat-elated state.



And then pitching (presentation of the idea). We had it at 12:55 (and at 12:40 it was necessary to be at the pitch zone), and code freeze was at 12:00. That is, we had about 30 minutes to prepare for the presentation. We prepared quite actively, and in the process of preparation we realized that the two slides should be in a different order, but fortunately they were adjacent and this did not hinder us much. Pitching went well, in a couple of places I tightened, but overall it was good. We were asked a couple of questions and we easily answered them, because were ready. Then we again had to wait, this time for 2 hours, because at 15:00 the finalists were announced. When the list appeared, we were depressed because we were not there, but depressed quite a bit, my mood personally still remained elevated, because even without a final I received a lot from this hackathon. This is where the hackathon story ends.

2 October

Today we were shown a table with points for pitching and I was pleasantly surprised. We ended up 4/23, which is pretty strong for the first time. To get to the finals we literally did not have enough, we can say that one completed feature was not enough.



Much on this hackathon was for me for the first time:

• The hackathon itself was for me the first full-fledged (before that it was only in Artificial Intelligence).
• This is the first time I have posted an application that is not ashamed to show and which I like.
• This is the first time I haven’t slept so much). For 55 hours I slept for 40 minutes.

PS: Hackathons are actually very interesting. Of course, we are still not close to winning them, but winning the track is more than a real goal. The atmosphere that is present sets you up for work and it’s even more pleasant to skip. Around like-minded people, you can chat with cool guys from different top companies (and not just about the program).



A reference to the application that we managed to write