A multi-agent AI system using Google AI Agent SDK and Gemini to analyze real-time route safety (crime, weather, lighting) and optimize navigation.

🐙 OZOO - Personal AI Assistant | 23 messaging platforms | Multi-agent routing | Transformed from 35% to 72% with security hardening & type safety

A portable `.claude/` bundle for Claude Code with a staged agent workflow, safety hooks, logging, and a strict verification loop. It also includes OpenClaw setup and bootstrap scripts for remote control, browser automation, multi-agent routing, and cross-session memory.

🛣️ Analyze and enhance route safety in real-time using a multi-agent AI system that assesses risks like crime, weather, and traffic.

LifeGuard AI is an intelligent, multi-agent system designed to simulate real-time accident risk assessment, emergency route planning, and critical decision-making for vehicles. The system demonstrates how autonomous agents can collaboratively analyze environmental data, predict risks, and make safety-critical decisions.

Lightweight AI agent orchestration engine — YAML workflows, multi-LLM routing (OpenAI/Anthropic), built-in agent patterns, safety layer, and observability. Production-ready with 185 tests.

A safety-aware multi-agent medical assistant that analyzes symptoms, verifies research, and delivers reliable health insights using RAG and intelligent query routing

Modular, deterministic oversight framework for multi-agent swarms: intent routing, behavioral ownership tracking, safety gates, auditable DAG trails. Skeleton / interfaces only (Jan–Feb 2026).

LifeLine is an AI-powered multi-agent disaster response system that detects hazards from images, analyzes real-time alerts, and guides people to safety with clear instructions and evacuation routes.

AI Disaster Guardian is an AI-powered multi-agent system designed to detect misinformation, deliver verified safety instructions, and offer shelter routing support during emergencies such as floods, fires, and earthquakes.

Agentic AI Multi-Agent Personalized Travel Planner is an intelligent system that uses multiple AI agents, vector memory, real-time APIs, and ML prediction to create personalized travel itineraries. It optimizes routes, budget, and bookings while adapting to weather, safety, and user preferences automatically.

A multi-agent mental health support system for students that combines safety-focused crisis triage, empathetic conversation, and mood tracking. Uses deterministic risk assessment tools and specialized agents (Crisis Specialist + Empathetic Listener) to route conversations safely and adapt responses over time.

An autonomous, multi-agent HR service desk. It uses a supervisor architecture to route employee queries between a RAG-powered policy librarian and a tool-calling clerk, featuring Human-in-the-Loop (HITL) safety controls.

Vision: Create a multi-agent counseling AI system that provides empathic, contextually appropriate mental health support while maintaining strict safety protocols and ethical boundaries. The system will leverage parameter-efficient fine-tuning techniques to specialize individual agents while using intelligent routing to coordinate responses.

An autonomous, multi-agent swarm control plane for mission-critical Smart City IoT and Defense logistics. Built with Agentic AI (LangGraph/CrewAI) for decision-making and Swarm Intelligence (ACO) for dynamic, self-healing routing under adversarial conditions. Features a Human-on-the-Loop (HOTL) veto protocol for operational safety.

Women’s Safety Companion is a non-emergency AI assistant that helps users plan safer trips with route guidance, transport tips, checklists, and contact reminders. Built with a multi-agent Planner-Worker-Evaluator design, it provides calm, practical travel advice without using external APIs.