Last weekend Peter, Neil and I attended HackTheHub, a machine learning based hackathon in Belfast. I was super excited to attend as machine learning is such a hot topic right now and it’s something I don’t really have much experience in. I had attended the pre-hack meet up which had some brilliant talks and gave me a real feel for the uses and possibilities of ML. It was a two day hackathon, the first day mainly about idea generation and building a prototype before showcasing and pitching your idea on the second day. With just over 30 teams signed up for the event, it was the biggest hackathon I have ever attended!
Of course, it all started with the most important thing — free food! Fuelling ourselves up with bacon butties and other good stuff, we had a long day ahead. After having some breakfast and walking around some of the sponsors stands (picking up a few awesome fidget spinners along the way), we grabbed a table in the main room and got ourselves set up with everything engineers need — post-its, whiteboards and Red Bull 😎
Ambitious.. but awesome?
We wanted to try out something different, not just the usual cleaning a big data set and using it to train a model. We wanted something quirky and innovative that would really stand out — after some deliberation we decided to try and build a chat bot that could generate a multi-platform application through just having a conversation. I know.. pretty ambitious for a days worth of coding, but we started with just trying to get a simple proof of concept where the user could pick between different application templates and Jenny would build it for them.
Once we got our proof of concept working we were all super buzzing about the idea, so we started to build it out more and add in new features. We fixed up Jenny’s dialog to make the conversation sound more human and added LUIS so that it could have natural language processing. This meant the user answering the questions wouldn’t have to use specific phrases e.g.
We wanted to try out something different, not just the usual cleaning a big data set and using it to train a model. We wanted something quirky and innovative that would really stand out — after some deliberation we decided to try and build a chat bot that could generate a multi-platform application through just having a conversation. I know.. pretty ambitious for a days worth of coding, but we started with just trying to get a simple proof of concept where the user could pick between different application templates and Jenny would build it for them.
Once we got our proof of concept working we were all super buzzing about the idea, so we started to build it out more and add in new features. We fixed up Jenny’s dialog to make the conversation sound more human and added LUIS so that it could have natural language processing. This meant the user answering the questions wouldn’t have to use specific phrases e.g.
“Jenny, I want you to create a new app for me”
“Generate an app”
will produce the same result!
We also added a feature that once the app was built, it would compress the files from the source code and send it back to the conversation with Jenny for the user to download.
Late on the Saturday evening with only 1 hour left to code, we also managed to get Jenny to customise the apps colour scheme once it was generated. The moment we saw the colour change happen was such a great moment for us, our heads were so fried from working all day and to see that little change was incredibly satisfying and also motivating. Crashing from the caffeine and one beer too many, we decided to call it a night on this achievement!
Demo Day
After a full day of coding on the Saturday, we all showed up the following morning and went straight for the coffee to top the caffeine levels back up. We quickly fixed a few things and got a bit of a demo prepared for the science fair. Each team sat at their tables and attendees, sponsors and judges walked around each of the tables to find on some more info on the idea and get a quick demo of what each team had achieved.
It was great to see a genuine interest in what we had manage to create in such a short period of time, our demos got a few ‘wows’ which definitely helped with the nerves of the final presentation!
After the science fair we had to give a final pitch, which was strictly 2 minutes and not a second longer. Since most people had seen Jenny the Jenerator perform during the science fair, we decided to focus on presenting the overall idea including the business side of it — we talked about the main benefits, how we could add more features in and how we could make money in the future.
Finally, after the judges had a quick discussion, we were delighted to have been placed 2nd overall — especially since there was a huge turnout of 30+ teams! It was really good to see our hard work pay off and getting some recognition from the judges and the sponsors. Time to celebrate!
So who would Jenny benefit?
We really thought that this chat bot would be very useful for small and medium enterprises to get a online presence (as a recent US survey revealed that more than 40% of SMEs still don’t have an online presence). These smaller companies potentially wouldn’t have the money to be able to hire a company to build them an app/website or just don’t have the technical skills themselves to do it. The fact that they can do it via a simple, natural conversation saves them both time and money.
Future Plans for Jenny?
We have decided as a team that it would be super interesting to see how far we could take Jenny, in terms of what features we could build and how much we really could customise the application for the end user. So stayed tuned and watch this space, we’ll keep you posted with our #JenUpdates!
Big thanks to HackTheHub for organising the super event as well as the fantastic sponsors including Bazaarvoice, Rapid7, Cayan and lots more! Thoroughly enjoyed by everyone and hope everyone took as much from it as I did!
