Summer 2024 Hackathon Winners Announced
The MS in Applied Data Science program and the Data Science Institute (DSI) recently wrapped up the exciting Summer 2024 Hackathon. This challenge was open to all MS in Applied Data Science students and alumni, allowing them to showcase their skills in a collaborative environment. Students formed teams of up to three and tackled the challenge using their preferred open-source pre-trained Large Language Models (LLMs) or Multimodal LLMs. The goal? Develop innovative retrieval, chatbot, or search systems to analyze provided data.
A rigorous evaluation process
The judging panel comprised a diverse group of experts, including MS in Applied Data Science professors, current Applied Data Science students, and the DSI’s Data Science Clinic Director. They evaluated each project based on a set of criteria, ensuring a thorough assessment.
AI-Cademic takes first place
The winning team, AI-Cademic, comprised Daniel Sa and Ariel Azria. They impressed the judges with their Gen AI application designed to enhance the analysis of various data types, including unstructured (text & image), semi-structured, and structured data from academic papers. Hackathon judge, and Associate Clinical Professor, Utku Pamuksuz, highlighted AI-Cademic’s exceptional performance on the sample queries, particularly those featuring complex and diverse scenarios.
That being said, the journey wasn’t without its challenges. Azria shared that the team spent considerable time mastering vector databases and information retrieval, “Before the hackathon, I had a basic understanding of retrieval-augmented generation (RAG) functionality. However, delving deeper into creating a precise vector database was a key learning experience.”
Azria further emphasized the importance of team dynamics, stating, “There were moments of frustration, especially with LangChain and embedding indexing. We went through the classic ‘napkin brainstorming’ phase before Daniel thankfully found a notepad. Cracking the indexing hurdle was a significant breakthrough. It allowed us to implement safeguards and prevent our model from generating nonsensical responses.”
Advice for future participants
When asked for advice for future participants, Sa stressed the importance of selecting a hackathon sub-topic (e.g.,academic papers) that aligns with team interests. He also emphasized building a team based on strong rapport and camaraderie. “Having reliable teammates who are eager to learn and enjoyable to work with is crucial,” Sa explained. “You can learn from each other freely and focus on the task without hesitation. When the project gets intense, maintaining a positive and communicative environment makes all the difference.”
Dharti Seagraves, hackathon judge and current MS in Applied Data Science student, recommends that students participate in hackathons: “The hackathons at UChicago in the Applied Data Science program are a great way to immerse yourself into real-world projects. It’s an amazing way to use the skills you’ve acquired throughout the program and also collaborate in teams and learn from one another and the faculty. The projects have a wide range of prompts and the ability to add your own touch and showcase your passions.”
Stay tuned for the next challenge!
This year’s hackathon showcased the talent and dedication of MS in Applied Data Science students and alumni. We look forward to seeing the innovative ideas that emerge in the next hackathon.