Local-First AI

The appeal of running AI without cloud dependencies.

The Problem with Cloud AI

Every time you send an image to a cloud service for processing, you’re trusting a third party with your data. For pet photos, this might be acceptable. But what about medical images, private documents, or proprietary business data?

Cloud AI also means:

• API costs that scale with usage
• Network latency that kills real-time applications
• Dependency on external services that can go down or change their policies

The Local Alternative

Local-first AI flips this model. Instead of sending data to the model, you bring the model to your data. This approach offers:

MemoMate

A web application that streamlines note-taking and summarization powered by AI. Upload documents, create AI chatbots from your notes, and generate diagrams via natural language.

Features

• Note Workspaces — Hierarchical tree-based file/folder organization
• Document Upload — Support for PDF, DOCX, and Excalidraw formats
• AI Chatbot Creation — RAG-based chatbots from uploaded notes
• Diagram Generation — Natural language to Mermaid syntax (flowcharts, Gantt charts)
• Whiteboard — Excalidraw-based collaborative drawing
• Command Bar — kmenu-powered quick actions
• Shareable Bot Links — Create bots from notes and share via unique URLs
• Admin Dashboard — Site metrics, token usage, and user engagement

Technical Highlights

• LangChain + LangGraph for RAG workflow orchestration
• Faiss vector store for similarity search
• OpenAI GPT-4o-mini for chat and diagram generation
• Excalidraw integration for whiteboard functionality
• Redux Toolkit for state management
• AWS EC2 deployment with GitHub Actions CI/CD

Tech Stack

Architecture

User uploads documents (PDF, DOCX)
         ↓
Document Parser (pdf-parse, mammoth)
         ↓
Text Splitter (RecursiveCharacterTextSplitter)
         ↓
Embedding Generator (OpenAI)
         ↓
Faiss Vector Store
         ↓
User chats → Retrieve relevant chunks → LLM generates response

Media Bot

An AI-powered media planning platform with a conversational interface for campaign optimization. Combines LLMs with mathematical optimization to maximize ROI across marketing channels.

Features

• Conversational AI Interface — Stage-based chat that gathers campaign requirements (budget, objectives, audience, channels, duration)
• Dataset Upload — Drag & drop CSV/Excel files with historical performance data
• Budget Optimization — SciPy SLSQP algorithm for optimal channel allocation
• ROI Analysis — Automatic calculation of cost efficiency and audience reach metrics
• Channel Support — Meta, Snapchat, Google, TikTok (expandable)
• Session Persistence — SQLite-based conversation history

Technical Highlights

• Together AI (Llama 3.1 70B) for natural language understanding and plan generation
• SciPy optimization for mathematical budget allocation
• Pandas for data processing and analysis
• Framer Motion for smooth UI animations
• Glassmorphic design with dark theme

Tech Stack

How It Works

1. Requirements Gathering — Chat collects budget, objectives, target audience
2. Data Analysis — Upload historical performance CSV/Excel
3. Optimization — SciPy finds optimal budget split across channels
4. Plan Generation — AI creates detailed channel-specific recommendations

Dynamic RAG

A sophisticated Retrieval-Augmented Generation system with multi-agent architecture for intelligent question-answering over custom knowledge bases.

Features

• Multi-Agent Architecture — Specialized agents working together:
  • QA Agent — Main agent for answering questions using the knowledge base
  • Personal Agent — For personal/general questions not requiring RAG
  • Validation Agent — Classifies queries as on-context, personal, bad, or unknown
  • Prompt Agent — Generates diverse query perspectives for comprehensive retrieval
  • Reranking Agent — Reorders retrieved chunks for better relevance
• Hybrid Search — Combines FAISS vector similarity with keyword matching
• Streaming Responses — Token-by-token streaming of LLM outputs
• Multi-language Support — Uses multilingual embeddings via Cohere
• Document Processing — PDF parsing with OCR support via Mistral

Technical Highlights

• LangChain + LangGraph for orchestrating the RAG pipeline
• FAISS vector store for similarity search
• Cohere embeddings for multilingual support
• SQLite for session persistence and analytics
• Configurable agents via JSON configuration files

Tech Stack

Architecture

ValidationAgent → classifies query type
       ↓
PromptAgent → generates specialized prompts
       ↓
QA Agent → retrieves chunks, reranks, generates answer
       ↓
StreamingResponse