On June 1, 2024, the tech world witnessed an extraordinary display of innovation and collaboration at the Upstage GenAI Hackathon. Sponsored by industry leaders MongoDB, LangChain, Together AI, MindsDB, and AI Camp, this event brought together brilliant minds eager to push the boundaries of technology. Among the standout achievements was the second-prize-winning project, NextMOVE, a Personalized Chess Training Assistant developed by Sadhana Arivoli, Aradhana Arivoli, Arivoli Tirouvingadame, and Jayaradha Natarajan.
Event Highlights
The hackathon kicked off with an enthusiastic welcome to participants, who were provided with a wealth of resources to ensure a productive and enjoyable experience. Key sponsors like MongoDB offered tools, credits, and extensive guides to help teams dive straight into their projects.
Key Resources
MongoDB Atlas: A unified platform combining operational and vector database capabilities, enabling powerful AI applications.
Credits: Participants received $50 Atlas credits to explore and build their applications.
Quickstarts and Code Samples: Comprehensive guides and GitHub repositories showcasing various integrations with MongoDB Atlas.
Spotlight on NextMOVE: The Future of Chess Training
Background
In the world of competitive chess, Grand Masters prepare for their matches with the help of a team that studies their opponents' past games. This preparation involves creating advanced, opponent-specific training plans. Recognizing the need for a more automated and precise approach, the team of Sadhana Arivoli, Aradhana Arivoli, Arivoli Tirouvingadame, and Jayaradha Natarajan set out to develop NextMOVE.
Objective
NextMOVE is designed to be a Personalized Chess Training Assistant. By uploading past games, users can receive AI-driven predictions of the next moves for each board position, tailored to their opponents' unique styles and tendencies.
Pre-requisites
The development of NextMOVE utilized the following tools and technologies:
Python: For scripting and development.
OpenAI: To integrate advanced AI capabilities using an API key.
MongoDB Atlas Vector Search: To create clusters and database instances for efficient data management.
Tech Stack
Python: For application development.
OpenAI: For integrating language models and AI functionalities.
MongoDB Atlas Search: Leveraging vector search capabilities for data retrieval.
LangChain: To manage and orchestrate the AI components.
LLM: Large Language Models for advanced AI predictions.
Streamlit: For creating an interactive and user-friendly interface.
Architecture
Development Process
Step 1: Setting Up MongoDB Atlas
The team began by creating an account on MongoDB Atlas and setting up their database cluster. MongoDB Atlas's vector search capability was crucial for managing and querying extensive chess game data.
Step 2: Integrating OpenAI
Next, they integrated OpenAI to utilize its powerful AI models. This enabled the prediction of next moves in a chess game by analyzing the opponent's past behavior and tendencies.
Step 3: Building the Application
Using Python and LangChain, the team developed the core functionalities of NextMOVE. They designed algorithms to process the chess game data, perform vector searches, and generate personalized training plans.
Step 4: Creating the Interface
Finally, they used Streamlit to build an interactive interface, making it easy for users to upload game data, view predictions, and interact with the training assistant.
Features of NextMOVE
Personalized Training Plans: Tailored plans based on the opponent's historical games.
Next Move Prediction: AI-driven predictions for any given board position.
Interactive Interface: A user-friendly interface for easy interaction and data management.
Judges:
Kasey Roh, Head of Biz US, Upstage
Raghu Banda, Senior Director of AI, SAP
Vivek Gangasani, Sr AI Architect, AWS
Anton Polishko, Director of AI, PolyAgent
Hrishikesh Garud, ML Research, Google
Mariane Bekker, Founder, Founders Bay
Prakul Agarwal, AI/ML Products, MongoDB
Christy Chen, Founding member, Opus Clip
Hackathon Winners Announced
The Upstage GenAI Hackathon concluded with the announcement of winners who showcased exceptional creativity and technical prowess. The NextMOVE team proudly took home the second prize for the best use of MongoDB, a testament to their innovative approach and dedication.
Congratulations to All Winners!
First Prize: GTA 6 – GenAI to assist the 911 call
Second Prize: NextMOVE by Sadhana Arivoli, Aradhana Arivoli, Arivoli Tirouvingadame, and Jayaradha Natarajan
Young Innovators in STEM
Sadhana Arivoli and Aradhana Arivoli, high school and middle school students respectively, brought exceptional enthusiasm and creativity to the hackathon. Their participation was especially appreciated by the organizers, highlighting the growing interest of young girls in STEM fields. Their achievement stands as an inspiring example for other girls aspiring to excel in engineering and technology.
Conclusion
The June 1 Upstage GenAI Hackathon was a remarkable event that highlighted the power of collaboration and innovation. We are excited to see how projects like NextMOVE will shape the future of AI-powered applications, providing new solutions and enhancing experiences in various fields.
For more detailed guides and code samples, explore their git repo:
Bình luận