Open Source AI Hackathon - Hackathon Judge
Power to AI builders hacking prototypes in 6 hours
I recently got the opportunity to judge the Open Source AI hackathon in Seattle. Over 13 teams participated and presented demos in hackathon. The Open Source AI is amazing community, led by Yujian Tang, is dedicated to gather tech professionals and encourage participation in Open Source AI community through various events such as presentations, workshops, and hackathons. This particular Open Source AI hackathon was hosted at the Microsoft Reactor, providing a platform for developers to spend entire day at hackathon and collaborate with peers.
I spent the day to engaging with participating teams, exploring their project motivations and discussing potential technological approaches for implementation. It was fun day interacting and understanding diverse perspective of individuals. The event began with a LlamaIndex tutorial and team formation, followed by a six-hour hacking time. Teams were given three minutes to present their final work within final hour of the event.
Here are my three takeaways from hackathon:
The gap between ideas and prototypes is closing really fast: I was amazed to see quality of prototypes teams have created within 6 hours. This includes ideation, planning, prototyping and deploying.
Every team’s focus on viability of idea developing into an actual startup: Most of the ideas were novel and pragmatic. Teams focused on hacking a viable prototype that has path to bring it to customers. Some teams also digged and presented metrics on market share and their potential business model.
Varied background brings diverse perspective: With participation from working professionals, each participant had unique insights to share and perspectives on the problem statement, exposing each individual to novel challenges that had not been previously considered.
Here are some of the ideas teams worked on:
InsightHire: This project assists recruiters and hiring managers by analyzing job descriptions and resumes to identify candidate strengths and weaknesses. It employs four agents—Job Description Agent, GitHub Agent, Resume Agent, and Scholar Agent—to generate comprehensive candidate profiles.
My Privacy Diary: An AI-powered privacy analytics tool that transforms your browsing data into meaningful insights about your digital behavior. Using advanced LLM technology (GPT-4o), it analyzes your browsing history to provide a comprehensive overview of your online activities.
[Winner] The trending-laughs project is designed to generate humorous content based on current news trends. It utilizes APIs from NewsAPI, OpenAI, and ElevenLabs to fetch news articles, process them for comedic potential, and produce audio outputs of the generated jokes.
[Winner] Presence.ai enables users to create a digital version of themselves by cloning their voice and crafting a personalized AI agent. Users upload personal information and voice recordings, allowing the system to generate an AI that responds to questions in their unique voice, based on the provided data.
RoamReader is an interactive chatbot designed to analyze and discuss a user's location history. By processing data from Google's Semantic Location History, it enables users to engage in conversations about their past movements and activities.
CharacterAiChatbot is a project that appears to involve AI-driven character interactions, as suggested by its name
[Winner] SocialServicesNavigation is a web application aimed at assisting users in navigating social service resources.
Apologies for other teams whose amazing work I cannot cover in this post.
And here is photo with one of the winners:
If you want to get connected with Open Source AI community, defintitely consider attending one of the upcoming events here: https://lu.ma/oss4ai