Celebrating Innovation with AWS and Supabase
A collaboration for GenAI hackathon by AWS and Supabase
Published Jul 31, 2024
Last Modified Aug 1, 2024
TL;DR: This is a post to celebrate the winners from the recent AWSupabase GenAI Hackathon and a short backstory on why this hackathon is amazing.
- co-authored with Donnie Prakoso (Principal Developer Advocate @ AWS), Loke Jun Kai (AI/ML Specialist SA @ AWS)
🥳 I love hackathons! I find hackathons a great way to work with others, explore new technologies, and build solutions to real-world problems.
The AWSupabase GenAI Hackathon, held from July 12th to 14th in Singapore, was a unique opportunity for developers and tech enthusiasts to collaborate, learn, and push the boundaries of what's possible with Generative AI. Organized by AWS and Supabase, the hackathon aimed to provide a platform for participants to experiment with cutting-edge technologies and develop innovative applications.
Here are some photos from the event, capturing the energy and excitement of the participants:
Before I continue the backstory of this hackathon, let's have a closer look at some really cool apps from the winners of this hackathon.
by Claire Seah, Mathias Beeh, Ong Yi Ting
Eventual.ly is an innovative platform designed to bridge the gap between online and offline interactions. Born out of the Temus AWS x Supabase Gen AI Hackathon in July 2023, it aims to combat loneliness and foster genuine connections in an increasingly digital world.
Demo video is available here.
The platform leverages advanced technologies, including React, Supabase, AWS Bedrock, and LlamaIndex, to create a seamless user experience. It uses AI-driven matching to connect individuals based on shared interests and personalities, facilitating the transition from virtual interactions to real-world friendships.
Eventual.ly enhances social engagement and supports local economies by encouraging users to participate in community-driven events and explore local venues. The platform creates personalized user profiles through interactive quizzes, ensuring more meaningful connections.
For local businesses, Eventual.ly offers a unique opportunity to connect with a targeted audience, host tailored events, and boost brand visibility without the need for extensive event planning or marketing resources. By addressing the global issue of loneliness and social isolation, particularly exacerbated by the recent pandemic, Eventual.ly aims to create a more connected and vibrant community, both online and offline.
by Zacchaeus Chok, Ye Kaihang
Demo video is available here.
Suparock Cloud Architect Agent is an innovative tool designed to simplify AWS architecture management. It leverages advanced AI technology to assist software engineers in understanding and optimizing their cloud environments.
The agent interfaces directly with AWS services through the CLI, offering suggestions based on the AWS Well-Architected Framework and recommending relevant AWS web services. It can autonomously execute AWS CLI commands and generate dynamic cloud architecture diagrams, providing an intuitive companion for cloud management.
Built on a robust tech stack, Suparock uses Amazon's Titan Text Embedding v2 for text processing, Anthropic's Claude 3 Sonnet as its core AI model, and Amazon Bedrock for serving foundation models. It employs Supabase for vector storage and LangChain for agent orchestration and retrieval augmented generation.
This tool addresses the challenges of complex command interfaces and the need for adherence to best practices in cloud architecture management. By reducing manual workload and ensuring architectures meet industry standards, Suparock streamlines cloud management processes.
by Gabriel Yang, William Wijaya
Demo video is available here.
Aspire AI is an innovative AI tutor powered by Large Language Models (LLMs). Its primary feature is an advanced image-to-speech function that transforms PowerPoint presentations into engaging audio explanations.
The system currently uses a soothing voice to deliver content. It is designed with the flexibility to incorporate other types of voices. This feature aims to make learning more engaging and reduce stress for users.
While initially developed as an educational tool for students, Aspire AI has broad applications. It can enhance workplace safety briefings, facilitate the review of missed presentations, and enable interactive Q&A sessions with PowerPoint content.
Aspire AI processes PowerPoint presentations of any size, allowing users to ask questions directly about the content. This technology extends the concept of document interaction beyond traditional PDFs, opening up a wealth of possibilities.
by Jiaming Wang, Xiuju Liu, Zihan Wang
Demo video is available here.
Fablette is the "Instagram for children's storybooks." It empowers parents to effortlessly create beautifully illustrated and educational stories using AI-powered story generation. This innovative platform transforms storytelling into a seamless, engaging, and creative experience for families.
by Aditya Ghosh, Manoj Sai
Demo video is available here.
JobSense is an AI-powered service tailored for tech professionals seeking to navigate the dynamic job market. It offers personalized insights on job trends, career feedback, and skill development opportunities.
The system employs daily web scraping of major job boards like Indeed, LinkedIn, and Glassdoor to collect up-to-date job data. This information is then formatted and stored efficiently in SupabaseDB.
Leveraging Amazon Bedrock, JobSense uses a fine-tuned AI model trained on the collected job data. This model analyses current market trends to provide personalized job recommendations and evaluates the value of various certifications and qualifications.
By delivering curated, actionable information, JobSense empowers tech professionals to make informed decisions about their career development and job prospects. It serves as a comprehensive tool for staying competitive in the rapidly evolving tech industry.
by Ong Min Chloe, Su Shwe Yee, Shannon Tan Yanxi, Timothy Chia
Demo video is available here.
SG Legal Guru is a Telegram bot designed to assist tenants with tenancy-related issues. It provides informal yet essential support by analysing text, image, and voice inputs to understand tenant concerns. The bot generates and utilizes embeddings to query relevant legal content and past chat history stored in Supabase databases. Using Bedrock LLM Claude, it delivers accurate and context-aware responses to help tenants navigate their rights and resolve issues effectively.
With around 25 teams participating and 18 submitted applications, I'm truly impressed with their ideas, the solutions they proposed, and how they built these solutions within 2 days with AWS and Supabase.
So, I reached out to my colleague and the lead organizer, Yee Fei Ooi 黄俞菲 , and had a discussion with her about this hackathon, why they delivered this event, and what and why Supabase.
Yee Fei We organized the AWSupabase GenAI Hackathon to help people learn and explore Generative AI technologies. With the growing popularity of Gen AI, many find it difficult to get started due to limited budgets. This event aimed to accelerate the learning process and provide a space for people to experiment with Gen AI tools.
By offering workshops and bringing people together, we wanted to make it easier for participants to overcome the initial challenges and begin their Gen AI journey, encouraging innovation and creativity in this exciting field.
Yee Fei Supabase is one of our key partners, and we found more and more builders in the region are attracted to Supabase because they're user-friendly and can help developers accelerate their application development more effectively and efficiently.
Donnie's note If you haven't tried Supabase, consider it to be the open-source alternative to Firebase. Which means, you can have your database (Postgres), authentication, storage, and of course, vector embeddings in one single service. Pretty cool, go check it out!
Yee Fei The hackathon process involved a one-day in-person event, comprising pre-hackathon workshops and hands-on hacking, followed by an additional 1.5 days for participants to continue working on their submissions from home before submitting them online. During the pre-hackathon activities, participants had the chance to attend AWS and Supabase workshops, equipping them with valuable knowledge and skills.
While the hackathon is over, it doesn't mean that you should stop learning or building your apps. Here's a list of additional resources for you to get started:
Plus, if you want to keep in touch with Supabase, reach out to growth@supabase.io. They'd love to have a conversation with you.
On behalf of developers, I'd like to express my sincerest gratitude to the organizers from AWS and Supabase for delivering this hackathon.
For you, winners and participants — Thank you for your dedication, collaboration, and more importantly, your passion to bring your ideas to solve various problems with GenAI technologies.
And for you, readers — Hopefully, this post has sparked your interest and motivated you to explore the world of Generative AI. There's a world of possibilities waiting to be unlocked.
See you at the next hackathon! 🤘🏻