AI-ML on ayyzenn

Agentic RAG: My Multi-Agent Setup with Gemini and ChromaDB

Wed, 29 Oct 2025 00:00:00 +0000

What I Built

Basic RAG retrieves docs and generates an answer. I wanted something smarter — a system that picks the right strategy and retries with heavier techniques when the first answer isn’t good enough.

This project uses three agents:

Router Agent — reads the question and checks answer quality
Basic Generator Agent — fast retrieval for simple questions
Advanced Generator Agent — uses heavier techniques for hard questions

The system tries the fast path first. If the answer is weak, it automatically switches to advanced methods like query decomposition, HyDE, and multi-query retrieval.

LlamaIndex + Text-to-SQL: What I Learned

Mon, 27 Oct 2025 00:00:00 +0000

What I Was Trying to Do

I had structured data in SQLite and didn’t want to write SQL for every question. LlamaIndex let me ask things like “which city has the most people?” in plain English — it figures out the query, runs it, and returns the answer.

Key capabilities include:

Translating natural language into SQL queries.
Converting unstructured context into structured insights.
Dynamically retrieving relevant schemas or tables for complex databases.

Note: Never expose full write-access databases to LLMs. Always use read-only credentials or sandboxed environments for safety.

RAG Text-to-SQL: Asking My Database Questions in Plain English

Sat, 25 Oct 2025 00:00:00 +0000

What I Built

I hate writing SQL for simple questions. This project lets me ask things like “how many customers are from Brazil?” in plain English — it generates the SQL, runs it, and gives me a readable answer. All local, using Ollama and ChromaDB.

Easy to use — ask questions in everyday language
Private — everything runs locally on your machine
Fast feedback — generates and runs SQL in seconds
Schema-aware — understands table names and relationships
Interactive — ask follow-up questions in a chat-style flow
Error handling — catches common issues like wrong table name casing

System Architecture

My RAG Text-to-SQL system consists of five core components:

Local RAG with Ollama and ChromaDB: Running AI Offline

Sun, 29 Jun 2025 00:00:00 +0000

Why I Built This

I wanted a chatbot that answers questions from my own documents — without sending data to a cloud API. This RAG setup runs fully offline with Ollama and ChromaDB on my machine.

Privacy — your data stays on your machine
No API costs — no usage fees or rate limits
Speed — no network delay for each query
Control — you choose what goes into the knowledge base
Offline use — works without internet after setup

Architecture Overview

My local RAG system consists of four main components:

YouTube Summarizer: What I Built with LangGraph and Gemini

Fri, 06 Jun 2025 00:00:00 +0000

What I Built

I wanted to paste a YouTube URL and get a summary — or ask questions about the video — without watching the whole thing. This project uses LangGraph for the workflow and Gemini for the actual summarizing and Q&A.

Key Features of Agentic Workflows

Modularity — each step can be built and tested on its own
State — data flows from step to step in a structured way
Reusability — the same steps can be reused in other projects
Autonomy — each step can decide what to do based on the input it receives

Popular Tools for Agentic Workflows

Tool	Description	Language	Best For
LangGraph	Graph-based AI workflow framework with state management.	Python	Custom step-by-step pipelines
CrewAI	Agent orchestration tool inspired by human teams and role delegation.	Python	Role-based agent collaboration
Autogen	Microsoft’s multi-agent framework for goal-oriented dialogue and tasks.	Python	Conversational multi-agent systems
LangChain	General framework for chaining LLMs with tools, memory, and logic.	Python	Broader LLM apps beyond just agents
AgentOps	Infra layer for deploying and monitoring agentic systems.	Platform	Scaling, testing, and managing agents

Why I Chose LangGraph

For this project, I chose LangGraph because:

Ollama + DeepSeek: How I Run LLMs Locally

Mon, 03 Feb 2025 00:00:00 +0000

Why I Use Ollama

I wanted to run LLMs on my own hardware without paying for API calls or sending data to the cloud. Ollama made that straightforward — install, pull a model, start chatting.

Install Ollama

Arch Linux

Update your system, then install Ollama:

sudo pacman -Syu
sudo pacman -S ollama

Enable and start the service:

sudo systemctl enable ollama
sudo systemctl start ollama
sudo systemctl status ollama

You should see active (running).