Upgrade Your Screen Time: Learn CS Fundamentals with Discite
Lazy mornings, late nights, and in between, we often find ourselves endlessly scrolling through the internet, browsing social media, catching up on the latest news, or checking notifications. Our phone usage is at an all-time high. Imagine transforming that screen time into something productive without altering our habits. Discite is a novel short-form learning app that aims to enrich your knowledge of the fundamentals of software engineering. Experience it for yourself by searching for Discite in the app store.
Exploring Discite
Discite is designed to offer bite-sized educational videos that enhance your learning journey. Recognizing the growing popularity of platforms that prioritize short-form content, such as YouTube Shorts, TikTok, and Instagram, we set out to bridge the gap between concise videos and educational content.
Our platform enables you to customize your learning by selecting the bite-sized educational content you prefer. Following the onboarding process, our sophisticated recommendation engine presents various learning options, making sure you keep advancing your skills in multiple categories while occasionally introducing fresh topics so you can discover other areas of potential interest.
Engaging with our platform is straightforward: you select a video and start your learning process by watching a concise clip selected from the more extensive educational video. For instance, in a video about Machine Learning, a single clip might address an atomic concept such as gradient descent or backpropagation. Our algorithm sequences the clips properly so you are only introduced to an idea after the concepts it builds on. We also have indicators showing how many video clips you are in a video. That way, you always know exactly how far you are in your learning path.
Our platform allows students to discover new topics and explore them more deeply. If a user is captivated by a topic, they can simply swipe sideways to access additional clips from the same video series. This makes it possible to grasp a subject in minutes instead of sitting through an hour-long video on platforms like YouTube. For those searching for different content, a downward swipe initiates a new series of clips on a related subject, allowing students to explore new interests.
Furthermore, our dashboard allows students to analyze their most engaged topics, offering a comprehensive overview of the content they’ve consumed. This feature not only fosters an interactive community by encouraging users to share their discoveries with fellow students but also provides a valuable tool for tracking learning progress and identifying new areas of interest. Our dashboard delivers an in-depth analysis of each student’s learning path, supporting them as they explore and deepen their passions.
Our Process
At the core of Discite’s development, we’ve harnessed a modern tech stack to create a seamless and efficient learning experience. Our application is powered by a Swift-based frontend to create an IOS app tailored for students. The backend API is managed by Node.js and MongoDB, ensuring scalability and effective handling of concurrent user requests. We’ve integrated AWS services for video streaming, providing reliable and high-quality video playback with zero buffering. Similarly, our personalized recommendation engine is built on a vector database, allowing sub-second querying and curation of video suggestions.
A unique aspect of our content strategy involves sourcing short-form content from YouTube’s extensive library of long-form videos. Initially, we aim to populate Discite’s database with machine learning-generated clips. This approach allows us to quickly amass diverse educational content covering a wide range of computer science topics. First, we use machine-learning models to match sections of our videos to topics of interest, creating time series data with topics. We then use our own change detection algorithm on this time series to identify optimal clipping points. This process results in a curated selection of short-form content that is both engaging and informative. As we scale, our vision extends beyond machine-curated content. Once we approach a critical mass of students, we aim to onboard content creators and allow them to upload content directly to the platform.
Our recommendation engine is also something worth highlighting. Inspired by YouTube and TikTok, our algorithm has three main components: affinity, complexity, and relationship. Each topic and subtopic has a similarity relationship with other topics, which is complemented by a video’s affinity for each topic and a video’s perceived complexity. Similarly, a user has a list of affinities (interest) to each topic and how complex they perceive each topic to be. By tracking a user’s interactions, such as watch time, likes, and dislikes across different videos and topics, we dynamically adjust their affinities and complexity scores toward topics. This personalized approach ensures our recommendations are relevant, engaging, and aligned with each learner’s educational journey.
Our Impact
In the digital age, where screen time often eclipses face-to-face interactions, Discite reimagines this time as an opportunity for learning and personal growth. We validated our platform by doing on-campus interviews with ~20 students. Many reported spending over 30 minutes daily on “doom scrolling” on TikTok, Instagram, or YouTube. Many of our beta testers reported feeling more fulfilled by our platform, which turns idle screen time into a step forward for education. Furthermore, many found our content more engaging and digestible than long-form content online, helping integrate Discite into an effortless part of daily life.
As we progress, the vision extends beyond just delivering content; it’s about cultivating a community of learners and content creators equally invested in the pursuit of knowledge. The impact of Discite is measured not only by the content consumed but also by the connections formed, the discussions sparked, and the curiosity ignited. Using our friends and sharing system, we want to make the community aspect of Discite a pivotal component of learning. In the long run, we want to foster a dynamic learning environment where users consume content and contribute to the collective knowledge pool. Our goal is to create a community of learners who help each other to do better.
Our Team
We are a team of Dartmouth students passionate about learning and building. Throughout our journey here at Dartmouth and recruiting for software engineering roles, we were frustrated with the existing system for learning new content in an engaging fashion. Furthermore, we saw firsthand the rise of short-form content and how engaging it could be, and it got us thinking.
Our team is composed of:
- Kevin, a ’24 interested in AI infra and optimizing scalable systems with previous experience at Amazon and TomTom.
- Amittai, a ’24 interested in deep learning and reinforcement learning, with previous experiences at Meta and the Robotics Lab at Dartmouth.
- Bansharee, a ’24 interested in the intersection between Human-Computer Interaction and Artificial Intelligence, with previous experience at Arista Networks and the Empower Lab at Dartmouth.
- Colton, a ’24 with interests in the crossover between AI and government, and their effects on future relations between nations, with previous experience as a web developer for Zamonos.
- Jessie, a ’24 interested in healthcare AI and sustainability software, with previous experiences at Nuance and GE.
- Yizhen, a ’24 hoping to bridge the gap between artificial intelligence and finance, with previous experiences including Google, Blackstone, and Bain Capital.
Conclusion
For anyone looking to enrich their knowledge in software engineering and beyond, Discite offers a new approach to learning that you will find practical and fun. By merging the convenience of short-form content with the depth of educational material, Discite promises to transform wasted screen time into moments of learning and discovery. With our advanced machine learning model, personalized content degenerated with our recommendation engine, and our passionate team at the helm, we hope Discite proves exciting and valuable for you. You can install Discite from the AppStore (available for iPhone and iPad). For more on how Discite works, feel free to check our website.
