We are delighted to welcome you to the Zenko Hackathon 2023 at the Politehnica University of Bucharest. This prestigious event brings together talented participants with the sole aim of collaborating to create innovative and revolutionary technological solutions.
- Introduction
- Project Goals
- Prizes
- Generative Agent Concepts
- Potential Project Focuses
- Project Datasets and Documents
- Rules and Guidelines
- Resources
- Submission
- FAQs
- Contact
This project offers an exciting opportunity for students to explore the world of artificial intelligence (AI) and collaborative agent-based simulations. Whether you're new to AI or an enthusiast, this project will provide you with hands-on experience in building AI-powered systems.
**Create an AI microsimulation **
The global vision is to create a microsimulation featuring AI agents that mimic the functionalities of ChatDev in an innovative way. Projects will be rated on 120 points.
The Zenko Challenge
Zenko Microsimulation has been designed with the following goals:
Collaborative Simulation: Explore the potential of AI agents working collaboratively in a simulated environment. GPT Integration: Showcase the integration of AI agents with powerful LLM models. Versatility Demonstration: Demonstrate the versatility of AI agents in various domains and tasks, powered by LLM.
Tasks and Points
- Create a virtual agent that handles the role of a director of customer relations (100 points)
- Integrate this virtual agent in a microsimulation environment (10 points)
- Add more virtual agents in this environment (10 points)
Please review the full breakdown of the hackathon tasks: Hackathon Tasks
- 1st Place: $1,000 and internship opportunity
- 2nd Place: $700 and potential internship opportunity
- 3rd Place: $500 and potential internship opportunity
The following document provides a comprehensive overview of the LLM (Large Language Model)-powered autonomous agents. Check the detailed sections:
- LLM Powered Autonomous Agents https://lilianweng.github.io/posts/2023-06-23-agent/
Participants are encouraged to select one of the suggested project focuses below or come up with a related innovative idea of their own. Each focus provides a concept, its applications, and the challenges associated with it.
Concept:
- Objective: Create an autonomous agent that assists in planning and optimizing events based on target demographics, current trends, and budget constraints.
- Functionality: Analyze previous event data, predict attendee preferences, and provide recommendations for event themes, venues, and schedules.
- Application: Conferences, weddings, corporate events, and festivals.
Challenges:
- Accurately predicting attendee preferences and adapting to changing trends.
- Integrating with various vendor systems for seamless event planning.
Concept:
- Objective: Implement an agent that autonomously promotes events to targeted audiences and engages potential attendees through personalized interactions.
- Functionality: Identify target demographics, design tailored promotional content, and distribute it across various platforms.
- Application: Event marketing for concerts, exhibitions, conferences, and more.
Challenges:
- Crafting compelling promotional content that resonates with diverse audiences.
- Real-time tracking of promotional campaign effectiveness and making necessary adjustments.
Concept:
- Objective: Build an agent focused on enhancing the guest experience at hospitality establishments like hotels, resorts, and restaurants.
- Functionality: Analyze guest feedback, preferences, and behaviors to recommend personalized services, amenities, or activities.
- Application: Hotels, resorts, restaurants, and other hospitality establishments.
Challenges:
- Ensuring guest privacy while collecting and analyzing data.
- Integrating with various hotel management systems for real-time service optimization.
**Vector Database Concepts
**Vector Database Documentation
https://python.langchain.com/docs/modules/data_connections
https://python.langchain.com/docs/integrations/vectorstores/chroma
**Script Examples
https://github.com/TestSavant/ingest_and_query/blob/main/ingest.py
https://github.com/TestSavant/ingest_and_query/blob/main/query.py
**Sample Client Docs
These documents are examples of the information needed for the agents to provide accurate and detailed event information
Hackathon registration closes on October 20st.
The hackathon starts on Saturday, October 21 at 1:00 PM UCT+3 and ends on Sunday, October 22 at 1:00 PM UTC +3.
After every project is presented, the jury will shortly select and announce the best projects.
Your project must be an MIT-licensed public GitHub repository.
Each submission will be cloned on Sunday, October 22 at 12:59 PM UTC+3 and will then be evaluated by the jury during the next days.No changes will be possible after this date.
Each group must communicate their GitHub repository URL through Discord.
An explicit README.md must be written and a 3-5 min presentation, in English, will be used by the jury to understand your project and its use.
Don’t forget that the delivery of this project consist of two things: your technical delivery and your pitch presentation. The jury will have only a limited time to evaluate the teams projects and this presentation will be really important in order to do so.
During the demonstration of your delivery, you will be judged on various points such as:
- The quality of the agent(s)
- The methodology that you used
- The ambitions you had regarding the final objective
- Clarity and objectivity in the presentation/documentation
OpenAI :
Langchain :
- Doc: https://python.langchain.com/docs/get_started/introduction
- Agent: https://python.langchain.com/docs/modules/agents
- GitHub: https://github.com/langchain-ai/langchain
HuggingFace :
- App: https://huggingface.co/
- Doc: https://huggingface.co/docs
- Transformers: https://huggingface.co/docs/transformers/en/index
- GitHub: https://github.com/huggingface
Agent Framework Examples :
AI multi-agent simulations
- https://github.com/yoheinakajima/babyagi
- https://github.com/camel-ai/camel
- https://github.com/OpenBMB/ChatDev
- https://github.com/geekan/MetaGPT
- https://github.com/OpenBMB/AgentVerse
Don’t forget that the delivery of this project consist of two things: your technical delivery and your pitch presentation. The jury will have only a limited time to evaluate the teams projects and this presentation will be really important in order to do so.
We are doing a project review session 6 hours after the hackathon starts (6:00 PM UTC+3 on Oct, 21th) ! Also mandatory 😉 The goal of this session is to do a technical follow-up, clarify and facilitate the development of your project.
If you have any questions:
-
Gather: > JOIN GATHER < Gather is a software mixing videoconferencing and movement in a 2D virtual universe. The hackathon is happening online on Gather, the Zenko teams & partners will be present for your questions and for presentations. https://www.youtube.com/watch?v=EK6CUMB_CNM
-
Discord: > JOIN DISCORD <
For any additional questions or assistance, please reach out to
- Zenko Hackathon support: Team Lead: Laur Neagu.