AI Agentic Workflows for data preparation and transformation
LangGraph agent that answers questions about sales data by calling analysis tools. Runs as a single ReAct agent or as a supervisor with two specialist agents. Tested with CI/CD.
A ReAct-pattern tool-using agent built from scratch in Python — reasons step by step, routes questions to a calculator or knowledge-base search tool, and cites its sources.
ReAct-style coding agent built from scratch in Python - custom action parsing, safety checks and approved bash execution.
AgentRunner — AI agent execution service built on the ReAct (Reasoning + Acting) pattern: reasons, decides and calls tools (Python, LangChain, Ollama)
A minimal, zero-bloat primer to learn agentic AI from scratch using local / Cogni-Brain inference.
Multi-agent ReAct pipeline for bias-free candidate evaluation — DEI compliance enforced mathematically via weighted Pydantic rubric.
AI-powered article generator for Dev.to and Medium
AI research agent that autonomously selects and calls tools (search, Wikipedia, arXiv, calculator) to answer questions with cited sources. Built with FastAPI + Groq.
Autonomous AI agent that operates a real web browser — perceives pages, reasons with an LLM, acts via Playwright, and extracts structured results. Built on the ReAct pattern.
Hands-on Agentic AI learning project — ReAct agents, memory systems, evals, and multi-agent architecture. Built as a structured AI PM curriculum.
Agentic AI sleep coach. Multi-agent system powered by CBT-I research.
An autonomous Python agent powered by Gemini that explores codebases, understands architecture, and implements product requirements using a strict 6-phase ReAct workflow. Built with native function calling and manual tool execution for maximum reliability.
Framework for tool-using LLM agents: ReAct loop, typed tool registry, and a deterministic offline backend.
Multi-agent system with planner, researcher, and writer agents collaborating via tool calling and shared-state orchestration
🤖 Intelligent, secure, and multilingual chatbot backend for Lorenzo Maiuri's website. Built with FastAPI, Gemini LLM, LlamaIndex, and MongoDB. Features session memory, tool-calling, and robust security
Bounded ReAct AI agent - tool allow-list, per-tool timeouts, AST-safe calculator, FastAPI, pytest, Docker
A step-by-step hands-on guide to building a tool-using AI Agent from scratch in Python with the Google Gemini API. Zero heavy frameworks, just core concepts.
Autonomous financial research agent implementing the ReAct pattern from scratch — multi-step tool use, two-tier memory (Redis + PostgreSQL), SSE streaming, and per-tool-call OpenTelemetry spans.
To associate your repository with the react-pattern topic, visit your repo's landing page and select "manage topics."