Repository Overview

AI Research Agent

Welcome to the repository! This document provides an overview of the projects and files contained within.

Projects and Directories

1. Kaggel-googleGenAi-porjects/

This directory contains projects related to Google's Generative AI, likely from a Kaggle competition or course.

Files:

• README.md: Project-specific documentation.
• day-1-evaluation-and-structured-output.ipynb: Jupyter Notebook for evaluating and structuring outputs from generative AI models.
• day-1-prompting.ipynb: Jupyter Notebook focusing on effective prompting techniques for generative AI models.
• day-2-document-q-a-with-rag.ipynb: Jupyter Notebook implementing a document question-answering system using Retrieval-Augmented Generation (RAG).
• day-2-embeddings-and-similarity-scores.ipynb: Jupyter Notebook exploring embeddings and similarity scores in the context of generative AI.
• day-3-building-an-agent-with-langgraph.ipynb: Jupyter Notebook demonstrating the creation of an AI agent using LangGraph.
• day-3-function-calling-with-the-gemini-api.ipynb: Jupyter Notebook showcasing function calling capabilities with the Gemini API.
• day-4-fine-tuning-a-custom-model.ipynb: Jupyter Notebook detailing the process of fine-tuning a custom generative AI model.
• day-4-google-search-grounding.ipynb: Jupyter Notebook utilizing Google Search to ground or enhance the outputs of a generative AI model.

Subdirectory:

• whitepapers/: Contains relevant whitepapers as PDF documents:
  • 22365_19_Agents_v8.pdf
  • 22365_3_Prompt Engineering_v7.pdf
  • Agents_Companion_v2 (3).pdf
  • neurips_evaluation.pdf
  • whitepaper_Foundational Large Language models & text generation_v2.pdf
  • whitepaper_emebddings_vectorstores_v2.pdf

2. composeToManifestScript/

This directory includes scripts for converting Docker Compose configurations to Kubernetes manifests.

Files:

• localScript.py: Python script for local execution of the conversion process.
• script.py: Main Python script to perform the conversion.

3. contanerizationScript/

This directory focuses on containerization, likely with scripts for generating Dockerfiles or managing containers.

Files:

• hostedLLMScript.py: Script designed to work with hosted Large Language Models (LLMs) for containerization tasks.
• localScript.py: Script for local containerization operations.

4. memory-chatbot/

Contains a chatbot implementation that incorporates memory capabilities to enable a choose your own adventure game.

File:

• memory-bot.py: Python script for the memory-enabled chatbot.

5. agents/research_agent/

This directory houses an implementation of a research agent, an AI agent designed to assist with research tasks.

Files:

• Dockerfile: Defines the Docker image for the research agent.
• cleanup.py: Script for cleaning up resources or data related to the agent.
• core.py: Core logic and functionality of the research agent.
• deployment_instructions.md: Instructions for deploying the research agent.
• docker-compose.yml: Docker Compose file for managing the agent's services and dependencies.
• main.py: Entry point or main application file for running the agent.
• requirements.txt: Lists the Python dependencies required by the agent.
• test_endpoints.py: Script for testing the agent's API endpoints.
• tools.py: Contains tools used by the research agent.

Subdirectory:

• templates/: Contains HTML templates for the agent's web interface.
  • index.html: Main page template.
  • result.html: Template for displaying research results.