Skip to main content

Agentic AI Framework

This module teaches the design, development, and deployment of autonomous AI agent frameworks using tools like Langchain, Crew AI, and Hugging Face. It covers everything from package publishing and API integration to building multimodal, multi-agent systems with CI/CD automation, monitoring, and public contribution pipelines.

Day 1-15:​

  • Pip package deployment, poetry package deployment

Deliverables​

  • Deploy your own pip package and submit the public github repo code and pip package link (Package should be deployed via CI/CD so that any updates after passing tests should auto-deploy to pip)
  • Poetry package management and deployment to pip

Day 16-30:​

  • Summary and Research report on:
    • Langchain Agents
    • Crew AI Agents
    • Pydantic AI Agents

Deliverables​

  • Detailed research internal document (PDF)
  • Public Blog tutorial with example code (github)
  • Sample AI Agent for the task of your choice via all 3 frameworks

Day 31-45:​

  • Open AI API testing
  • Gemini API testing
  • Hugging Face API testing
  • Fast API testing

Day 46-60:​

  • Architecture development for Prodigal AI Agents Framework
    • Should cater to Multi Agent systems
    • AI Agents Auxiliary and support architecture
      • Input Validation
      • Output Validation
      • Hierarchy rules

Deliverables​

  • Draw.io complete design for the AI Agents
  • Detailed doc for development philosophy and milestones

Day 61-75:​

  • Prodigal AI Agents Framework Code structuring & Issues/Contribution Guidelines

Deliverables​

  • Initial code structure and basic package
  • Delivery milestones via Issue Trackers (Issue Bot Management)
  • Sample AI Agent Demo

Day 76-90:​

  • Complex Agents Deployment
    • Reference: Ben AI (combination of AI agents and make.com Automation)
    • Auto Blogging Agent/ Auto SEO Agent

Deliverables​

  • Live Agent Automation performing Blog Automation / SEO (based on choice, first preference is Blog Automation)

Day 91-105:​

  • Addition of support for multiple LLM models and custom models

Deliverables​

  • Support for OPEN AI, Gemini Pro, Hugging Face models, and other LLM models
  • Create proper guidelines for integration of custom LLM agent

Day 106-120:​

  • Planning for Agents log monitoring and visualization
    • Clone features for LangSmith
    • Clone features for LangGraph

Deliverables​

  • Feature list research document for implementation of the log monitoring and debugging of agents
  • LangGraph based visual element and branch based support

Day 121-135:​

  • Implementation of log monitoring and visualization Phase 1

Deliverables​

  • Code integration for monitoring and visualization

Day 136-150:​

  • Implementation of log monitoring and visualization Phase 2

Deliverables​

  • Code integration for monitoring and visualization

Day 151-165:​

  • Plan Improvement of Multimodal Agents
  • Utils requirements to support multimodal agents: text, image and videos

Deliverables​

  • Detailed research doc for the planning of the requirements for improvements for support for multi-agents systems

Day 166-180:​

  • Implementation of public contribution pipeline to allow easier public contribution
  • First time issue tags, levels of template issues to allow easier commits
  • Moderator management for commits

Deliverables​

  • Implementation of public contribution issues and moderation systems including CI/CD issue bot tracker

Tech Stack​

  • Python
  • PEP style coding
  • pip package development
  • poetry package management
  • Langchain
  • RAG
  • Advanced RAG
  • Cache RAG
  • Crew AI
  • Microsoft Autogen
  • Hugging Face
  • LLM Integration
  • LangSmith
  • LangGraph
  • AWS
  • Digital Ocean