Vision-grounded RAG agent for financial documents, built on a self-finetuned ColPali and LangGraph.

Jeu liant les MATHS et PYTHON

Blokus est un jeu multijoueur en 2D dont le but est d'encadrer plus de point que ses adversaires.

Multi-tool AI orchestration that researches its own success criteria before acting, validates them after, and earns autonomy from track record. Orthogonal to single-agent frameworks.

GPT Flow: Simplifying tasks with human-like AI, inspired by autogpt.

🥈 2nd place (Agent-only) MLSys 2026 Track B — Multi-Agent LLM Pipeline for Memory-Constrained Tensor DAG Scheduling. Neurosymbolic split: LLM for combinatorial structure (op fusion), deterministic solver for constraints ([w,h,k], dependency, latency). 7.77× aggregate speedup, 88× best case.

🔁 REFLEX — Reflexive Self-Improvement for AI Agents. Log failures → analyze → propose → apply. Part of the GENESIS AGI stack.

Interactive Streamlit application that benchmarks direct prompting, chain-of-thought, self-consistency, tree-of-thought and reflexion techniques across OpenAI GPT-3.5 and Groq Gemma-9B-IT.

A self-reflective LLM agent with tools, memory, and reasoning built using LangChain + ReAct + Reflexion. Modular FastAPI backend + Streamlit UI.

매일 쌓이는 메일을 빠르게 파악하고, 중요한 정보를 놓치지 않도록 돕는 LLM Agent 기반 Chrome Extension 서비스입니다.

A self-correcting SQL agent using LangGraph + Streamlit that learns from errors, reflects on mistakes, and automatically fixes SQL queries through multi-attempt reasoning.

MYJarvis - Personal AI Operating System built on Claude Code Max. Self-improving, multi-channel, 24/7 personal AI with zero per-message cost.

Adversarial and reflexion critique executor for iterative AI task refinement (ADR 0005)

An autonomous coding agent with persistent trauma memory. Implements "Loop 10" to save execution failures to a "Scar File," ensuring the agent never makes the same syntax or library error twice in a session.

🚨 AI-powered emergency response simulation combining Mesa ABM with advanced LLM reasoning strategies (ReAct, Reflexion, Plan-Execute) for optimized crisis management and multi-agent coordination

🐿Madus: A production-ready multi-agent framework that understands text, figures, and tables in complex PDFs to answer natural language questions

kairos picks the right LLM technique for any task by querying a curated wiki of agent patterns - RAG, ReAct, Reflexion, and 17 more. Karpathy LLM Wiki + rule-based selector + 3 working runners. Zero API keys (routes through llm-mcp).

Provide persistent, outcome-weighted memory for AI trading agents to improve decision-making across sessions with the TradeMemory Protocol.