Datathon

Six months ago AI NI held our inaugural Hackathon; an event bringing together companies and community to pull off Northern Irelands largest Hackathon. With the aim for over 250 attendees over two days and two locations, we are back again with another Hackathon!

Our Datathon focuses on the data aspect in relation to AI systems and developing the understanding that a machine learning pipeline is only as good as the data you give it. The attendees had access to a range of open source and sponsor provided datasets with the aim to produce a data driven solution to a societal problem. With skills development at the core of all AI NI events, this Datathon provided the attendees with access to mentors, data, computing platforms, soft skills development and technical workshops; meaning every skill level of attendee could be supported and catered for.

Our Aims

Skills and community development is always in the forefront of AI NI events and the team wanted to build on the previous Hackathon (you can read that blog here) to make sure we were offering as many learning opportunities as possible. With our challenges the goal was to ensure the company and AI NI proposed challenges were of societal benefit and addressed the need of the larger public; with this in mind the challenges have a local focus but could utilise global data to demonstrate a solution.

Further to this we wanted to offer teams the opportunity to develop their ideas further and bring them to life with financial backing post-event. To do this Techstart were brought in as part of our judging panel with teams able to submit their entries to their program offering £3000 initial support for making their idea into a possible business.

Day One

We rolled out the same pattern as our previous Hackathon (if its not broken don’t fix it) with attendees signing in, picking up their badges, free t-shirts and grabbing coffee before heading up to our opening keynote. Our two Academy leads, Chloe and Jake, delivered the keynote speech in our Belfast and Derry locations giving the details on the event agenda, workshops, catering and of course what everyone wants to know, the prizes that were up for grabs. Our Diamond sponsor, AquaQ, were then invited up for their keynote giving an overview of their company and their proposed challenge.

The Datathon had 4 challenges proposed to tackle various areas with AI solutions:

Data analysis for improving clean and renewable energy by AquaQ
Unlearning bias in recruitment by Smashfly
Solving the climate crisis by Dell
Detecting skin cancer by Peltarion
Mental health awareness or AI for Good by AI NI

With all the information laid out and the briefs for challenges set, the hackers were all set to go!

Event Schedule

We had phenomenal demand at our last Hackathon for workshops and mentoring support, so we brought back both of these aspects and worked to up the number of mentors and volunteers making sure the team could engage with as many attendees as possible.

Starting off the first day our own Mark Cunningham provided a workshop on Python101 for AI NI, to get all the less confident attendees up to scratch with basic coding. This workshop would provide the fundamentals and basis for all the sponsor workshops later in the day.

Following this we had our Platinum sponsor workshop from Aqua Q with Matt Doherty covering data visualisations, the need for developing data understanding and interesting projects in which AquaQ had utilised visualisations to enable insights for clients.

After lunch we had our wonderful Smashfly workshop, lead by Caiomhe Carbery. This workshop was titled “Dirty Data or Biased Data?” giving the attendees and insight into the Smashfly challenge, a quick run through of some machine learning concepts and hints for possible solutions.

At this point the workshop track split into two segments with the next workshop for Techstart on how their funding and prize would work, running simultaneously with a drop in clinic. The drop in was a way for mentors to sit with teams who were struggling and offer them support and guidance without being under time pressure.

The second parallel workshops were from MCS Group and Dell on “Accelerating AI: The Computation Cost” looking into how CPU’s from different manufacturers compare in offloading computational workload to FPGA’s or GPU’s. Our recruitment partner MCS Group simultaneously to this were running a CV Clinic where attendees could drop in for CV tips and to talk about available opportunities.

With that jam packed schedule for our attendees we finished the day off with some pizza and drinks for dinner, last minute hacking and a good nights rest to get ready for day two!

Day Two

On the morning of day two the teams had time to complete their last minute touches to their projects and get their pitch videos ready. Having so many attendees and teams meant that long pitches would have taken us hours, so to combat this the teams had 2 minutes to tell us why their project was the best! One of our biggest points of feedback from our previous event was that other teams missed seeing the winner solutions. This time round while the judging happened we held a showing of the videos so other teams could see the pitches and projects that other teams created.

Of course we didn’t have just our top prizes, we gave out prizes for the best meme, thank you vouchers to all our volunteers, finalists who made it to the last stage of judging, a highly commended prize for a team that just missed the rankings and the best AI NI challenge. Alongside the main Datathon prizes, Techstart awarded their grant of £10,000 and support for one team to take their idea further from the event. Lastly we had our final ranking:

The AI NI team always walk away from these events with a huge feeling of accomplishment, not because AI NI pulled off another large Hackathon with the backing of companies, but because the attendees have a brilliant time learning, upskilling and networking.

These events are largely supported by our sponsors and it is down to them that we are able to offer datasets, full catering, prizes and the resources; therefore thank you to AquaQ, Smashfly, Dell, MCS Group, Techstart, Peltarion and Slice; as well as our site hosts QUB and UU.