QHacks 2023 - Pitch Perfect (2nd Place)

QHacks 2023 was an incredible experience where my team competed against 188 talented engineering students from all across Ontario. For this hackathon, my team had 36 hours to create a product for the theme "Designing a Digital World". We created "Pitch Perfect" which generates slide decks using Artificial Intelligence. The program goes beyond just generating graphics and texts, as it also creates a script that fits perfectly into your presentation length. It completes your idea and flushes out as many details as possible without compromising your starting point.

Out of 47 teams we placed 2nd, and on this page I will describe our design process over the 36 hours. If you want to see our code (and try it for yourself) you can check out our GitHub (https://github.com/Sanbroh/QHacks2023).

---

How We Built It

We used OpenAI to generate outputs based on 3 inputs given by the user. The data was then processed to be outputted on pitch deck templates (created on Figma). The code was made to be modular for future expansions, and UI/UX design was considered throughout the entire process.

The front end involved HTML/CSS and JavaScript code in implementing the UI, as well as functions to support the core features of our completed MVP. The backend used Flask and Python to power advanced API calls from and to OpenAI, and was used to process data effectively for frontend deployment and rendering.

Besides the slide deck templates and UI/UX design/implementation, everything else was created using AI. The following two pictures show our homepage where the user puts in the 3 prompts, and an example slide that was generated. In the example slide, all of the content and the script to the right was generated using AI.

Challenges We Ran Into 

​There were many challenges in terms of frontend, backend, and design. The frontend integration with backend took a very long time, and there were many iterative design cycles to complete the frontend design. The backend involved many optimization and data processing issues, where we found it difficult to get the most ideal results for our test cases.

​For design, we expanded our scope many times to include new ideas into our design files, so that they can be implemented in the future. However, this definitely drove up the creation time, and pushed our team to the limit to deliver a fully-completed initial product as our MVP.
​
I was personally responsible for product design and the work to the right was our final prototype. On the left side I created the skeleton slides for our HTLM front end, and the right side was our product pages.

Below are a couple pictures from our pitch to the judges!

​

Next Steps

Our immediate next step involves creating more templates to satisfy more target audiences and designing a sound and fair monetization system. After that, we hope to optimize our design with a landing page and other support pages for our brand, and redesign some of the code to be more efficient.
​
We also hope to incorporate a database into our program, such as Firebase and MongoDB, to storage data for our users and allow membership on the website. Many features would be impossible without a data storage system, so we aim to get that done as soon as possible.

Lastly, we would love to sit down and do more research. This includes both user research and market research. To make a successful product, we are very interested to understand more of the industry and our target audiences, and hopefully discover new insights and rooms to improve.
​
This was a fantastic project, and we hope you like it as much as we do!

One cool part of the project was we pitched to the judges using slides (and scripts) that our product generated!