-
-
Figma: Landing page
-
Figma: Personal dashboard
-
Figma: Create project page
-
Figma: Project Kanban
-
Figma: Project files
-
Figma: Project analysis
-
GIF
Registering an account
-
GIF
Creating a project
-
GIF
Creating a Kanban tile
-
GIF
Moving Kanban tile
-
GIF
Moving and editing a Kanban tile
-
GIF
Inspecting ticket descriptions
-
GIF
Melbourne University assignment about designing and creating a music arcade came in Java
-
GIF
Monash University assignment about making a parallel string searching algorithm based on shared memory
-
GIF
Accessing project files
-
GIF
Accessing the virtual meeting room
Inspiration
We wanted to address one of the biggest issues that we face in group projects – people not pulling their weight. It’s easy to attribute this to skill differences, but often, people just struggle taking that first step. There’s confusion about what needs to be done and when.
There is much anxiety faced when taking in longer project specifications, and not knowing how to break it down into digestible tasks. We need something that can extract to-do items from the spec. Not knowing which item to prioritise or start with? Something needs to exist where it will allocate priorities for each ticket task, so you know what needs to be done first.
We also know how challenging and confusing it can get, coordinating all the different communication platforms and collaboration tools to use with your team. Sweet aims to mitigate that by providing a one-stop-shop for all the tools necessary for your next group project.
We always hope that we get allocated an organised team member, who can collect everyone’s contact details to share files with, and have a high-level overview of the project progression. Instead of hoping that we luck out in team allocations, our product solves this issue and acts as a constant and centralised product suite for group collaboration in projects. In this day and age, it would also be unheard of to not integrate some AI in a hackathon…
What it does
Welcome to Sweet, your go-to hub for all things group projects. Sweet will help effortlessly steer you through the entire project journey from specification to the final delivery.
We differentiate ourselves from competitors like Jira, Notion and Trello, with a loveable AI teammate, ‘Freckle’, who leverages generative AI to maximise your group’s projects efficiency and accountability. On top of that, while competitor platforms are separate, Sweet brings them all together in a convenient location, where you can hold or plan meetings, store files, create tickets and analyse your progress - AI integration is just the cherry on top! With reminder messages to keep you accountable on ticket completion and checkpoint prioritisation, Freckle is an organised and valuable team member in any group project.
Our features include:
- AI-generated Kanban board with project tickets and their corresponding checkpoints and priority
- Integrated group messaging with our AI teammate (Freckle) with ticket generating capabilities
- Tailored analysis of project progression and group collaboration
- Built-in suite of collaboration tools such as a shared file space, group scheduling and a 3D virtual meeting-space
How we built it
In the process of constructing our product, Sweet, we began by conceptualising a name that encapsulates our commitment to offering a comprehensive suite of collaboration tools. Drawing inspiration from the concept of a candy wrapper, we envisioned Sweet as a neatly packaged solution that houses all the essential tools in a cohesive and user-friendly manner. This metaphor extends to our chosen colour palette, where the richness of the purple, reminiscent of Cadbury, adds a visually pleasing touch to the overall user interface.
Our approach extended to the creation of Freckle, an AI teammate designed to be approachable and fun, embodying a persona that enhances user interaction. During the design phase, Figma played a crucial role, enabling us to visualise and refine wireframes, ensuring an intuitive and engaging user experience.
Turning to the technological foundations of Sweet, we adhered to industry standards:
- ReactJS + Typescript for Frontend (UI)
- AWS Amplify for frontend hosting
- 3JS for 3D modelling
- ExpressJS + Typescript for Backend
- MySQL Database
- AWS S3 for file storage
- OpenAI ChatGPT for LLM queries
- PDF2json for scanning and parsing assignment spec files
A pivotal aspect of our technological stack involved the integration of OpenAI ChatGPT for natural language processing queries. Prior to inputting data into our large language model (LLM), we implemented a strategic preprocessing step. Through a Hugging Face tokenizer, our data undergoes refinement, optimising API credit usage and ensuring more accurate and efficient results.
Overall, the careful consideration of every element in our development process has culminated in the creation of Sweet—a powerful, intuitive, and engaging project management solution.
Challenges we ran into
Our initial idea focused on a product suite for group collaboration, however upon realising the lack of importance/related issues to solve, we thought about what we desired in any group project work. Pivoting and committing to a focus on utilising AI to maximise productivity and project completion was challenging, particularly given we were having this change of mind at 5pm on the second day of the hackathon. Less than 24 hours were left and we still were not committed to any one hackathon project idea.
Taking a step back, we realised the importance of prioritising what a consumer would need. We brought ourselves together and channelled the very values of the product Sweet which we wished to inject into our own group collaboration and morale. Perhaps if we had our AI teammate Freckle to help us, we would have had a clearer plan of action and task allocation.
In the hopes of creating an effective group project experience - centralised, with the seamless integration of an AI teammate who catalyses group productivity, we deliberated on features we decided were priority. Doing this, we once again reflected on our personal experiences working on university projects and hackathons to whittle our ideas down to the few we decided were most pertinent. This was an important step in our process, considering our timeframe and our intention in creating a no-frills, fleshed-out experience with tools users would actually use.
Accomplishments that we're proud of
It actually works. Our calls to generative AI can actually create meaningful tickets which are relevant to the inputted project specifications. In an age where AI is everywhere, our project utilises this new generation of technology to add to group efficiency and collaboration, values which we see as being key in university projects and when these group members eventually enter a workforce. We take pride in the fact that this is a product which we would genuinely use in our daily lives, and can see far-reaching, tangible use cases where our product makes a difference in people’s collaborative skills.
Our UI design also creates a lively and friendly environment for collaboration, through our colour scheme and particularly with our AI being wrapped in a cute candy-themed character, Freckle. We can truly see our character being everyone’s favourite teammate, programmed to have a can-do attitude whilst keeping everyone accountable for their contributions. Tailored for university students, maybe even at a hackathon :), our product fosters a team-spirit not only through the actual functionality of our service, but through how we face our consumers too.
On a more team-focused level, we are proud of our team’s learnings in how we collaborate too. Through friendly verbal sparring about our hackathon idea, we now have a group state where all of our contributors are able to effectively work on what is meaningful to them whilst also learning new skills; at this very moment, 2:31 pm on Sunday with 2 hours and 28 minutes left until submission, all our pieces from the back-end API calls, to the front-end engineers implementing the UI built by our designers, we can feel the excitement and team-spirit as we see our idea come to life. Freckle is approaching the finish line!
What we learned
We can breakdown our most important lessons in roughly 3 categories
Having a northern star: An opportunity such as UniHack opens the door to an endless plethora of possibilities. This can get overwhelming quickly and teams can find themselves lost. Initially, our team was scattered in our priorities and unclear in our vision, but realising the importance of coming together with our ideas and forming a really clear northern star provided us the guide to our project progression. We formed a central problem statement, which resonated with our personal experiences and designed an application revolving around this. Soon we found ourselves creating a product we would genuinely find useful and can truly stand behind.
Combining individual skills and being open to learning: In any team, avoiding stepping on each other's toes and letting people work their craft is key. Our team brought together a heavy leaning towards back-end software engineering, and rather than being focused on what our background skills lent us to, we realised that being open to filling skill gaps is key to being able to deliver a product to consumers. Across our team, we have newly born Figma experts and React programmers who learnt how to make our entire service front from 2x speed youtube tutorials on the way to the Melbourne Hub. Within the short time frame of the hackathon, our mindsets regarding what we individually can contribute have been broadened, with a meaningful and functional product to show for it.
Passion is key: We couldn’t have worked through tiny misalignments in our Figma diagram and all nighters trying to fix our build, without passion for our craft and the very service we created in the past 2.5 days. Our discussions about product prioritisation and actually making a start on our service was fuelled by a desire to make something meaningful in the realm of group productivity - something which we could pour our heart, brains, and souls into. Every single member was willing to put our everything into this project because of our belief in what we were making.
Reflecting and comparing this to prior projects, hackathons and even industry-work, we leave this hackathon project realising the importance of working on what we believe in. And how that makes all the difference in being able to push out a product which caters to what people need and want, with a happy smile on our faces as we do so.
What's next for Sweet
In our pursuit of enhancing Sweet, we are exploring the capability to incorporate specialised tools that cater to specific domain needs, such as integrating GitHub for programming projects. This not only adds functionality but also aligns with our goal of providing a one-stop hub for all resources, making it more convenient for users to access everything they need in a centralised platform.
To improve user experience, we are considering the implementation of a scheduling integration in a similar way to ‘When2Meet’, that establishes a consistent time for meetings. This feature aims to simplify the process of allocating meeting times, ensuring seamless coordination among team members.
We'd also love to integrate video conferencing, a documentation editor and Figma integration, for an even smoother production process. An additional key feature that we'd wield with Freckle AI would be transcribing live meetings and automatically generating summarised minutes after such meetings.
Drawing inspiration from successful platforms like Duolingo, we intend to develop the distinct persona behind Sweet. Much like Duolingo's engaging approach, we envision Freckle, our platform's character, evolving dynamically, possibly losing its "sprinkles" as project tickets remain uncompleted. This approach not only adds a unique touch to the service but also serves as a visual representation of task completion and project progress, injecting personality into our project management experience.
Built With
- amazon-web-services
- amplify
- express.js
- figma
- mysql
- openai
- react
- s3
- tailwind
- three.js
- typeorm
- typescript




Log in or sign up for Devpost to join the conversation.