Ai Hackerspace - July 5

Jul 5 / rUv

AI Hackerspace Live:

This week’s AI Hackerspace showcased a variety of insightful presentations and discussions. Highlights included Nazim Ghorash's LinkedIn tool integrated with Microsoft Teams, Bence Csernak's hackathon project on facial reconstruction using stable diffusion and GPT-4.0, and John Petty's work on cognitive and modular AI platforms. The session featured detailed technical demonstrations, engaging lightning talks, and vibrant networking, emphasizing innovation, practical applications, and community engagement in the AI space.

Summary Template


Introduction

The session was opened by Reuven Cohen, setting a collaborative and enthusiastic tone. He welcomed everyone and highlighted the scheduled lightning talks, mentioning the technical difficulties due to Microsoft and Zoom updates. Reuven emphasized the importance of staying adaptable and maintaining engagement despite the disruptions.


Event

  • Type: Online Meeting
  • Date: July 5, 2024
  • Platform: Zoom


Participants

  • Reuven Cohen (rUv): Host
  • Nazim Ghorash: Presenter
  • Bence Csernak: Presenter
  • John Petty: Presenter
  • Kelsey Cole: Participant
  • Other Participants: Various professionals including Nick, Gideon Marken, Bradley Ross, and more.



Summary

00:00:00 - 00:00:23

  • Speaker: Reuven Cohen
  • Content: Introduction to the meeting.
  • Details: Reuven welcomed participants, mentioning the scheduled lightning talks and the technical difficulties due to Microsoft and Zoom updates.


00:00:23 - 00:03:24

  • Speaker: Reuven Cohen
  • Content: Overview of the meeting agenda.
  • Details: Announced lightning talks, featuring Nazim Ghorash's LinkedIn tool, Bron Phil Walker's SEO insights, and John Petty's AI platform.


00:03:24 - 00:03:47

  • Speaker: Reuven Cohen
  • Content: Mentioned technical difficulties.
  • Details: Explained that both Windows and Zoom decided to update simultaneously, causing disruptions.


Presentations


Nazim Ghorash's Presentation

00:03:47 - 00:04:23
  • Speaker: Nazim Ghorash
  • Content: Introduction to his LinkedIn tool.
  • Details: Discussed the concept of a LinkedIn tool integrated with Microsoft Teams, which won a hackathon in the healthcare space.

00:04:23 - 00:08:54
  • Speaker: Nazim Ghorash
  • Content: Demonstration of the LinkedIn tool.
  • Details:
    • Functionality: The tool generates LinkedIn posts based on inputs such as company details, reference posts, and desired tone of voice.
    • User Interface: Showcased the UI with examples of posts for companies like DoorDash and HNBC.
    • Technical Integration: Utilizes the Microsoft Bot Framework to deploy across platforms like Teams, Slack, Outlook, etc.
    • Demo: Presented the process of generating a LinkedIn post using LangChain code and bot framework emulator.

00:08:54 - 00:10:47
  • Speaker: Reuven Cohen
  • Content: Follow-up questions and discussion.
  • Details: Reuven asked about the similarities with AutoGen Studio and integration challenges. Nazim clarified differences, mentioning that his tool focuses more on integration within the Microsoft ecosystem. He also mentioned plans to write an article on the tool.


Bence Csernak's Presentation

00:10:47 - 00:11:23
  • Speaker: Bence Csernak
  • Content: Introduction to his hackathon project.
  • Details: Explained the hackathon context and his project on facial reconstruction using stable diffusion and GPT-4.0.

00:11:23 - 00:19:54
  • Speaker: Bence Csernak
  • Content: Detailed presentation on the project.
  • Details:
    • Hackathon Context: The project was developed during a 48-hour hackathon organized by a New York City company.
    • Objective: To use AI for facial reconstruction, especially for burn victims and those requiring plastic surgery.
    • Technical Approach: Combined stable diffusion for image generation and GPT-4.0 for vision and contextual understanding.
    • Demonstration: Showed the app's functionality, including uploading photos, getting surgery assessments, and comparing before and after images.
    • Team Dynamics: The team was randomly assigned but worked well together, leveraging their skills in AI and healthcare.

00:19:54 - 00:24:57
  • Speaker: Participants
  • Content: Q&A and feedback.
  • Details: Participants praised the project and discussed its implications in healthcare and cosmetic industries. Bence mentioned future steps involving real surgeons and medical professionals to improve the tool's accuracy and applicability.


John Petty's Presentation

00:24:57 - 00:25:23
  • Speaker: John Petty
  • Content: Introduction to his project.
  • Details: Discussed his team's work on cognitive and modular AI platforms, focusing on data handling and contextualization.

00:25:23 - 00:34:03
  • Speaker: John Petty
  • Content: Detailed presentation on polymorphic data handling.
  • Details:
    • Objective: To improve how data is processed and contextualized for AI models.
    • Technical Approach: Utilized Jinja templates and JSON for structuring data, enhancing the LLM's understanding and processing capabilities.
    • Process: Described the steps of deconstructing and reconfiguring JSON, adding intelligence and context iteratively.
    • Use Case: Demonstrated an example of a cognitive node handling complex information dynamically, contextualizing data based on user roles and interactions.
    • Security: Mentioned incorporating security components to handle sensitive data and ensure contextual awareness.

00:34:03 - 00:39:57
  • Speaker: Reuven Cohen and Participants
  • Content: Q&A and technical discussion.
  • Details: Reuven inquired about the use of function calling and libraries like Marshmallow. John explained their approach, including the use of JSON5 and Beautiful Soup for data handling. Discussed potential applications and enhancements using NetworkX for complex data relationships.


Networking and Community Discussion

00:39:57 - 00:42:38

  • Speaker: Reuven Cohen and Kelsey Cole
  • Content: LinkedIn group migration discussion.
  • Details: Kelsey suggested moving the LinkedIn group to a page for better visibility. Reuven agreed and encouraged participants to share their LinkedIn profiles.


00:42:38 - 00:43:44

  • Speaker: Various Participants
  • Content: Sharing LinkedIn profiles.
  • Details: Participants shared their LinkedIn profiles for networking purposes.


Chat and Social Media Links

Key Chat Exchanges

  • Microsoft Update Day: Mentioned by multiple participants, acknowledging the disruptions caused.
  • Kelsey Cole's Masterclass Topic: Focused on AI use cases in consumer brands.
  • LinkedIn Community Link: Kelsey shared a link to the Artificial Inspiration community on LinkedIn.
  • Networking: Participants shared their LinkedIn profiles for better connectivity and collaboration.


Social Media and Other Links


Conclusion and Action Points

00:43:44 - 00:44:57

  • Speaker: Reuven Cohen
  • Content: Wrap-up and future plans.
  • Details: Thanked participants for their presentations and participation. Discussed plans for future sessions and collaborations.


00:44:57 - 00:46:30

  • Speaker: Reuven Cohen and Kelsey Cole
  • Content: Marketing and community support.
  • Details: Kelsey offered to help with marketing and community engagement. Reuven expressed appreciation and plans to follow up.


Summary and Action Points

  • Action Points:
    1. Reuven Cohen to work on migrating the LinkedIn group to a page for better visibility.
    2. Participants encouraged to share and connect on LinkedIn.
    3. Kelsey Cole to assist Reuven Cohen with community and marketing efforts.
    4. Follow-up on the articles and code demos presented during the meeting.


    Upcoming Ai HackerSpace Live 

    Created with