AI_ENGINEER.init()

Ido Degani

AI Software Engineer & Data Scientist | LLMs, Agents, Infra

5 years building AI Pipelines, LLMs & agents

I'm deeply passionate about artificial intelligence and its transformative potential to reshape how we solve complex problems. From developing sophisticated multi-agent systems to architecting scalable LLM infrastructures, I thrive on pushing the boundaries of what's possible with AI engineering.

When I'm not building the next generation of AI systems, you'll find me diving deep into research papers, experimenting with cutting-edge architectures, or contributing to open-source projects that advance the field forward.

LangChain MCP LLMs RAG Vector DBs Agentic workflows Git Multi-Agent Python PyTorch TensorFlow LlamaIndex OpenAI API Hugging Face AWS Docker FastAPI

Featured Projects

A collection of AI engineering projects showcasing LLM implementations, agent architectures, and scalable infrastructure solutions.

filter_by:

Fact-Checked AI Analyst: Agentic RAG Workflow

Advanced multi-agent workflow combining RAG techniques with semantic caching for efficient analysis of financial documents. Features intelligent agent orchestration for retrieval, fact-checking, and response generation with embedding reuse optimization. Implements token-aware processing and similarity-based prompt caching. Achieved 80% reduction in processing time for repeated queries while maintaining answer consistency on Apple's 10-K reports.

Open in GitHub

LangGraph Semantic Caching RAG Multi-Agent Systems

Multi-Agent Research Swarm System

Sophisticated autonomous research system using LangGraph to orchestrate specialized AI agents that collaborate to produce comprehensive reports. Features advanced state management, conditional workflow logic, and dynamic task distribution. Agents operate with distinct expertise areas including planning, research, synthesis, and report generation. Integrates Tavily API for real-time data gathering and implements supervisor pattern for efficient scaling.

Open in GitHub

LangGraph Multi-Agent Autonomous Systems Research Automation

Advanced RAG System with Pinecone and LangGraph Agent Workflow

Advanced RAG system combining Pinecone vector database with LangGraph's multi-agent workflow for NASA Systems Engineering Handbook processing. Features cyclic agent architecture (retriever, generator, refiner, reviewer), optimized document chunking, and semantic caching. Gradio-based chat interface enables intelligent semantic search and context-aware response generation with OpenAI embeddings integration.

Open in GitHub

Pinecone LangGraph RAG Multi-Agent Systems OpenAI Gradio Vector-DB Token Optimization

RAG-Powered API Documentation Assistant

Intelligent documentation assistant that automatically searches, retrieves, and generates executable code from API documentation. Implements complete RAG pipeline with web crawling, semantic chunking, and vector storage using ChromaDB. Features code generation with GPT-3.5/4 and provides cited explanations. Reduces documentation lookup time from manual searching to instant code generation with source attribution.

Open in GitHub

RAG ChromaDB Web Scraping Code Generation

AI Sales Intelligence Platform

Multi-agent system automating B2B prospect research using 5 specialized LangChain agents orchestrated by a supervisor. Features web scraping, RAG with vector storage for pitch patterns, and performance observability. Reduced research time from 30 minutes to 2 minutes with 3x higher response rates.

Open in GitHub

LangChain Multi-Agent RAG Vector-DB Web-Scraping

Customer Support Automation System

Intelligent multi-agent system for e-commerce support handling 200+ daily tickets. Built with 4 specialized agents using MCP integration for CSV data access and RAG pipeline for knowledge base queries. Achieved 100% query resolution with sub-second response times.

Open in GitHub

LangChain MCP-Integration RAG ChromaDB Multi-Agent

Multi-Agent API Documentation Assistant

Real-time API learning assistant using coordinated multi-agent workflow. Features web search, documentation extraction, RAG-powered retrieval, and automated code generation with explanations. Processes any API documentation into working Python examples with setup instructions and common pitfalls.

Open in GitHub

Multi-Agent RAG Code-Generation API-Integration Real-Time-Search