Blog

Abstracta Hackathon 2024

Curious about what a hackathon involves? In this article, we tell you everything about it: its meaning, potential impact, and what you should consider when organizing such a transformative event for your team. We dive deep into the lessons and results that came from our 2024 edition, along with all our team proposals.

Collage of the five teams who participated in the Hackathon

Everything moves at such an astonishing speed that it can sometimes be hard to keep up. Additionally, each company, organization, team, and even each individual, has its own motivations and needs, making it impossible to cover everything.

At Abstracta, we believe deeply in the power of experimentation. We allow ourselves to make mistakes to continuously learn. We embrace them as stepping stones to progress, innovate, create, and form networks along this path. 

Being able to pause, shift focus, and take the time to immerse ourselves in new technologies collectively allows us to gain knowledge and learn by doing, with a real and practical focus.

This is why in 2023, we started organizing hackathons to drive innovation and learning, enhancing the quality of our work. This has helped us consolidate our team and complement each other across technical and support areas through diverse perspectives and skills.

In this article, we highlight our 2024 hackathon, focusing on Generative Artificial Intelligence and the creation of copilots.

Looking for a quality partner to boost your business with Generative Artificial Intelligence? Explore our AI Development & Copilots Service!

What is a Hackathon?

A hackathon is a collaborative event where professionals from different fields, each with diverse knowledge and skills, come together to work in teams. They develop innovative tech solutions to predefined challenges in a short time.

These events can last from a few hours to several days. Participants, usually in teams, dive into an intense process of brainstorming, coding, and prototyping sessions.

By bringing together people with varied perspectives, hackathons are often highly motivating spaces that spark disruptive ideas and out-of-the-box solutions. They foster an environment of creativity and experimentation where teamwork is elevated to its highest expression.

Read our Generative Artificial Intelligence Guide for beginners and beyond in this article!

What to Consider When Planning a Hackathon?

Team working at the Montevideo Office

Organizing a hackathon is a challenge full of possibilities and variables. It involves making a series of decisions that will define not only the participant experience but also the event’s potential impact.

In this section, we share key items to consider on this path. Additionally, we tell you how we defined these points in our latest edition of Abstracta, with the aim of sharing and illustrating each step.

1.Theme Selection

Gen AI Copilots

The theme definition is crucial as it will guide all activities and projects during the event. It’s important to choose a topic that’s relevant and interesting to your community while allowing for the exploration of new ideas and innovative solutions.

At Abstracta, this year we chose “Generative AI Copilots,” a field with impactful possibilities and current challenges.

2.Training and Pre-Hackathon Tools

People working at the Salto Office

Providing training sessions and tools before the hackathon can add significant value. These sessions help level the knowledge and offer access to platforms for testing even before the event.

At Abstracta, we organized pre-hackathon training sessions, offered consulting spaces, and provided tools for experimentation for four weeks before the event. This fostered conscious participation and motivated teamwork.

3. Venue Specification

People working ath the Montevideo Office

Choosing between an in-person, remote, or hybrid event will significantly impact logistics, necessary technology, and interaction methods.

At Abstracta, we opted for a hybrid approach. We conducted the hackathon both in our offices and remotely to maximize inclusivity and flexibility.

4. Team-Building

How and when teams are formed can greatly influence the event’s dynamics. Considering the diversity of technical knowledge and experience may enrich the creative process.

At Abstracta, we prioritized diverse profiles. We formed groups of 4 to 6 people, some entirely in-person and others remote. We ensured that all teams included members from both technical and support areas.

We decided to organize it 20 days before the event. This allowed teams to start working on their ideas and solutions before the synchronous phase, arriving better prepared.

5. Proposal Presentation

Defining how projects will be presented is key to enabling a fair and organized process. It’s important to determine everything possible in advance, from presentation times to spaces for questions.

In our hackathon, each team had 10 minutes to present their proposal and an additional 10 minutes for Q&A.

6. Evaluation Criteria

Establishing clear and objective evaluation criteria is essential for organization and transparency. Sharing these criteria in advance with all participants helps them prepare as best as possible.

Our criteria at Abstracta reflected aspects like innovation, feasibility, and presentation quality. We shared them as follows:

Evaluation Criteria

7. Awards

Winning team with their prize
CompassBot was the winning proposal of the Hackathon 2024

Prizes are not the main goal of the event but rather the learning achieved along the way, the creation of ideas, networking, and bond strengthening. Nonetheless, prizes are nice gestures that help recognize effort.

At Abstracta, we decided to give a branded thermos to all participants and special backpacks to the winning teams as a token of appreciation for their dedication and hard work.

Communication

Some of the participants of the Hackathon

Effective communication is crucial for informing and maintaining participant enthusiasm. Focusing on this area helps the message to reach the desired audience.

At Abstracta, we utilized various communication spaces and platforms to inform about the event, convey its relevance, activities, and dynamics, and promote full participation. We did this through our monthly meetings called Abstracta Talks, internal team meetings, and daily chat channels, starting several months in advance.

9. Agenda

Hackathon 2024 Agenda

Planning a detailed agenda helps organize time efficiently and lets participants know what to expect at each moment. This avoids uncertainties and allows each team to manage their time optimally.

Our agenda included a collective breakfast, sessions dedicated to teamwork, and final presentations. There was a break for the jury to analyze the projects and a feedback session to highlight each proposal’s and team’s strengths.

10. Support Teams

Fabián Baptista, Alejandra Viglietti, Roger Abelenda  and Matías Reina
Fabián Baptista, Alejandra Viglietti, Roger Abelenda, and Matías Reina

Having dedicated support teams for different event aspects is vital to address team needs. Important points include support for logistics, technology, and consultations.

At Abstracta, we formed a specific team for the general event logistics, which involved decisions about the venue, materials, food, and more. We also had a team of judges and consultants to assist with technical and business data.

11. Work Modality

Hybrid work modality

Deciding whether the work will be synchronous, asynchronous, or a combination impacts both the event’s flexibility and accessibility. It’s crucial to determine this based on the context, the team composition, and the event objectives.

For our 2024 hackathon, we chose a mixed approach. In 2023, it was entirely in-person. This new way of conducting it allowed for remote participation from those who otherwise couldn’t join, as well as from those who could experience it in person. It enabled us to start working on the project before the hackathon day.

These carefully considered decisions boosted us to create a hackathon that met our innovation and learning objectives while providing a rich and accessible experience for all participants.

Pre-Hackathon Training and Tools

Providing training is essential to enable technical team members to experiment with tools before the event. This also allows support team members to either experiment if they wish or understand the scope and possibilities offered by the available tools, helping them propose their ideas effectively.

In this section, we share all about the training we offered before our 2024 hackathon.

A month before the event, we started offering workshops and opened weekly consultation spaces. This allowed all team members, regardless of their participation in the hackathon, to start learning and experimenting with generative artificial intelligence.

We also provided valuable tools for the processes: ChatGPT accounts for different teams, N8N, and PrivateGPT, a platform developed by our R&D team, based on Microsoft Azure, allowing access to multiple LLMs, including GPT-4.

Workshop #1: Exploring AI Copilots and Assistants

Processes and Tasks Impacted by AI Chart

First, we presented the activities and processes currently impacted by generative AI within Abstracta. Our goal was to improve both the quality of our products and services and our productivity.

Then, we focused on the differences between copilots and assistants. We reviewed examples of typical use cases such as interacting with documents, APIs, and databases. We analyzed the characteristics of OpenAI’s GPTs and assistants. To do this, we conducted demos of internal and third-party examples on each topic. Finally, we shared a concrete methodology for designing a copilot for a specific task or role.

Workshop No. 2: How to Create Assistants and Copilots

Genexus example on how to interact with APIs

We delved into building AI systems. We discussed Retrieval Augmented Generation (RAG) and developed a GPT for querying documents and external APIs.

You can find more information about RAG and related topics in this article.

We also presented the n8n tool, after sharing its tutorial in advance. We explored how it allows us to quickly generate prototypes. We also discussed Langflow and its ease in prototyping and chaining prompts and different tools to solve complex problems.

Workshop #3: Browser Copilot

Browser Copilot Demo

We reviewed the internal architecture and Langhain framework. We gave an overview of our open-source tool, Browser Copilot, and presented the copilot we created for Bantotal.

We shared information about LangChain, LangSmith, AzureFastapi Tiangolo, devbox, Keycloak, Vite plugin Web extension, Web search con ChatGPT API by Open AI, and Bing Web Search API by Microsoft.

Results of Our 2024 Hackathon

Part of the team members who participated in the Hackathon

We are thrilled with everything that emerged from the event: new ideas, innovative tools, and strengthened bonds through synergistic work.

We had 5 teams of 4 to 6 people, who tackled the challenge with dedication and commitment.

Below we share a summary of each proposal that emerged from our 2024 hackathon:

Team “LinkSense”

Team “LinkSense”

Focused on the growth area, this group aimed to create a copilot for generating messages en masse when contacting potential clients. They sought to personalize these messages by considering the client’s profile and context to build connections more effectively.

Their intention is to improve the meeting ratio, which is currently 1 meeting for every 1000 messages sent, and to increase the company’s opportunities for client expansion.

Team “Compass Bot”

Team “Compass Bot”

To explain the potential of CompassBot, it’s necessary to briefly mention what Compass is.

It is a tool that allows periodic and continuous performance monitoring and self-monitoring. It consists of a series of rubrics, where each person at Abstracta performs their evaluation autonomously and is also evaluated by their leader, with scores ranging from 1 to 5. 

Then, in a shared meeting, they discuss each point in detail and agree on the evaluation.

At Abstracta, we strongly believe that our capabilities result from our efforts. We understand that measuring and improving our performance can help us achieve better results.

However, we face the challenge that these instances consume a lot of time or, sometimes, don’t reach the level of reflection we seek. Giving constructive feedback can also be challenging.

This team’s proposal involves creating an innovative AI-based assistant that helps with self-evaluations and evaluations. The assistant ensures comprehensive reflection before the shared instance by considering everything necessary when suggesting a score for each rubric.

The tool asks specific questions in a private environment, allowing for reflective evaluations in a trustful setting. After the iteration, it generates a perceived and recommended score, which is saved in a file for final review. This helps to improve criteria and optimize the decision-making process.

This file will be useful in reducing the time these processes take while deepening all relevant human aspects of the journey, making the experience meaningful for the growth of each team member.

Example of Iterations
Example of Iterations

Connected to Open AI APIs, the assistant will be available through Slack via a chatbot. It will send a reminder when it’s time to perform the evaluation and start a dialogue considering all Compass rubrics.

Team “MinutAIzer”

Team “MinutAIzer”

This team presented an innovative solution to current communication challenges in the global job market. 

Nowadays, we often need to attend multiple online meetings, which is sometimes impossible. Yet, we need the information shared in those meetings to understand the interactions and make better decisions.

The team’s proposal is MinutAIzer: a system based on N8N with a centralized repository containing meeting transcripts. This assistant allows different roles to interact with the stored information.

Team “MinutAIzer”

Copilot Features:

  • Generate meeting minutes
  • Answer questions about one or several meetings
  • Search for agreements made
  • Analyze participants’ behaviors
  • Offer insights during live meetings

For example, you can ask questions about topics of interest, inquire about someone’s stance on a specific topic, and access personalized minutes.

Team “High Impact”

Team “High Impact”

The High Impact team presented User Story Builder, an AI assistant for automatically generating user stories (HUs.) This arises from the need for clear, structured HUs in projects to deliver software continuously.

In today’s world, projects require constant changes, making it hard to adapt HUs to changes at the needed speed.

How will the User Story Builder solve this? Using a Python agent for Browser Copilot extension and concatenating prompts in the system, the assistant will perform the following steps:

  • Thoroughly analyze the provided documentation
  • Identify system functionalities
  • Generate an HU for each flow of each functionality, specifying evaluation criteria for each

Currently, analysts manually perform this cycle in about three weeks. With User Story Builder, this would be reduced to a day, as the assistant would instantly create the user stories. The team can then validate and make necessary modifications.

We believe this assistant will significantly impact productivity and the quality of work processes within agile methodologies. It will allow us to analyze complex documentation crucial for development and testing teams.

Team “KBM Abstracta”

Team “KBM Abstracta”

KBM presents a practical, agile solution for Abstracta employees to obtain information about internal regulations applied to their specific cases instantly.

This copilot centralizes all company policies and regulations, from licenses to benefits and compensations. It’s possible to interact with this information easily through natural language queries via a Slack chatbot.

This helps resolve doubts immediately without human interaction or lengthy document reading, leaving personalized consultations for necessary cases.

Having centralized, easily accessible information can save time, increase productivity, and improve the query experience by providing answers precisely when needed.

Below, we share an example of a potential interaction with the chatbot:

example of a potential interaction with the chatbot

Learnings

Team Presentation

Abstracta’s 2024 Hackathon allowed us to live enriching experiences and cultivate many learnings. These collaborative instances reflect our work style and the synergy within our teams.

Here are some of the learnings we achieved along the way:

  • The Value of Facing Challenges

We confirmed once again that the most authentic learning arises when directly facing challenges. It’s in those moments we need to analyze precisely and focus on solutions. We realized it’s hard to measure the full potential of our learning.

  • The Power of Synergistic Work

Despite the inherent challenges of forming teams in such a short period, the 2024 Hackathon demonstrated that achieving extraordinary synergy is possible. 

This experience reaffirmed our belief in the power of collaboration and people’s ability to unite and create innovative solutions when common goals are set in an organized manner.

  • Consolidation of Our Team

We experienced firsthand how these spaces strengthen our bonds and allow us to connect with people from other areas we don’t usually interact with. This is enriching for team consolidation and understanding our potential more deeply.

  • Challenges in Pre-Event Preparation

For various reasons, several registered participants couldn’t attend the event, requiring team reorganization. 

We observed that while the idea of pre-forming teams and asynchronous work seemed promising, it didn’t work as well in practice as expected, with some notable exceptions. This invites us to rethink our preparation strategies for future editions.

Each of these learnings drives us to keep innovating in how we approach, plan, and execute events like the hackathon. It strengthens our ability to adapt quickly to changes and fosters a creative and collaborative work environment.

We hope this guide on how we conducted our hackathon inspires your organization. Do you have other ideas that you think would be interesting to implement when organizing such an event? Feel free to share your contributions with our team.

Looking for a quality partner? 

Explore our case studies! We are one of the most reliable companies in software quality engineering. Discover our solutions and contact us to discuss how we can help you grow your business.

Abstracta Ilustration

Follow us on Linkedin & X to be part of our community!

445 / 446