https://github.com/bharath-tars/agenticai

This repo is solely for study purpose of AgenticAI
https://github.com/bharath-tars/agenticai

Last synced: 27 days ago
JSON representation

This repo is solely for study purpose of AgenticAI

• Host: GitHub
• URL: https://github.com/bharath-tars/agenticai
• Owner: Bharath-tars
• License: gpl-3.0
• Created: 2024-12-30T06:37:30.000Z (27 days ago)
• Default Branch: main
• Last Pushed: 2024-12-30T06:41:37.000Z (27 days ago)
• Last Synced: 2024-12-30T07:30:36.661Z (27 days ago)
• Size: 0 Bytes
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# AgenticAI

This repo is solely for study purpose of

AGENTIC AI - Notes by Bharath

# Important points:

Started in Mid 2024 also called AI Agent, Multi AI Agents

More than GenAI

The diff between GenAI and agentic AI is:

GenAi -> create content -> LLM Model

Query ->(prompt) -> LLm -> Output {Generating content}

Also called prompt finetuning

RAG - > Retrieve relevant content that is upto date by giving access to an llm to an external source where data is stored in vectors like ChromeDB(Langchain is used)

Agentic AI -> Autonomous AI Systems / Agents (perform their own tasks)

They have a goal to achieve

All the AI systems work independently to achieve the goal

LangGraph (Agentic AI framework), langflow, Phidata

Solve Complex workflow

Can finetune themselves to perform much more better

Summary

Genai when access to external source makes workflow complex, in agentic AI goal is based on business outcome

Example:

Custom Bot -> Agentic AI -> interacts with many things

Query: Hey i have 1000$ Buy some stocks and sell it in 5 days that can give 50% profit.

Query: Compare tesla and nvdia stock and suggest best one to purchase

Here each workflow is a separate task and each task is handled by an AI separately. These all work together simultaneously  to achieve the goal and then give the output.

Open source framework 

###### AutoGen(Microsoft) -> https://github.com/microsoft/autogen

###### LangFlow(No code tool for Agentic AI) -> Dragdrop -> End code will we given -> Later can be deployed - > https://www.langflow.org/

###### LangGraph(More complex) -> https://www.langchain.com/langgraph

###### PhiData - > https://www.phidata.com/

###### CrewAI -> https://www.crewai.com

# The Influence of AI in 2025: Trends you should not miss

Agentic AI

AI Infrastructure -> Gpus as PAAS, Inference compute engine(Fast responses)

Large Models (LLm, LVM) -> Billion to Trillion parameters

AI Edge Devices for Gen AI -> Jetson Nano oron

Small Models(Phi3) -> OpenSource less para for fine tuning

Security AI -> Regulations

Lot of Open Source Models

# Prerequisites

Python

LangChain 

# Reference Links:

###### What is Agentic AI? Important For GEN AI In 2025 -> https://youtu.be/xOS0BhhdUbo?si=eiRK95aGk-sp_Hpw

###### Agentic AI Is The Future- GEN AI Trends In 2025 -> https://youtu.be/9SD95yskEfU?si=DxJLuS3AwkIylf_H

###### Building Agentic AI Free Course For Everyone -> https://youtu.be/qR3HWsMFfZA?si=PJ0FZllSd5RLU3KB

###### Detailed Prerequisites To Start Learning Agentic AI With Free Videos And Materials -> https://youtu.be/Qs_j5wRbVr8?si=oKDGRbilnhfgG_v0

###### Build your own Agentic AI Applciation using PhiData -> https://youtu.be/74SnvbQYgx8?si=-7aHhmGjxZHLWlre

###### Building Multi Agentic AI RAG using VectorDataBases -> https://youtu.be/4MTtfTZnH5Y?si=q-w_BVZ2joJMfDLy

# Phidata Docs

###### Docs url: https://docs.phidata.com/

###### Models: https://docs.phidata.com/models

###### Agents: https://docs.phidata.com/agents

###### Knowledge(External source or domain specific): https://docs.phidata.comj/knowledge

###### Storage: https://docs.phidata.com/storage (Formats)

###### VectorDbs: https://docs.phidata.com/vectordb 

###### Chunking: https://docs.phidata.com/chunking 

###### Embeddings: https://docs.phidata.com/embeddings 

###### Tools: https://docs.phidata.com/tools