Google DeepMind

Top Builders

Explore the top contributors showcasing the highest number of app submissions within our community.

Google DeepMind

Google DeepMind is a world-renowned artificial intelligence research laboratory, formed from the merger of DeepMind and Google's AI division. It stands at the forefront of AI innovation, responsible for groundbreaking advancements, including the development of the Gemini series of multimodal AI models and the Gemma open-model family. DeepMind's mission is to solve intelligence to advance science and benefit humanity.

General
Author	Google DeepMind
Release Date	2010 (DeepMind founding)
Website	https://deepmind.google/
Technology Type	AI Research Organization

Key Research Areas and Achievements

Reinforcement Learning: Pioneering work in reinforcement learning, including AlphaGo, which defeated world champions in Go.
Large Language Models: Development of advanced LLMs, contributing to the Gemini and Gemma model families.
Scientific Discovery: Application of AI to accelerate scientific research, such as AlphaFold for protein structure prediction.
Safety and Ethics: Dedicated research into AI safety, ethics, and responsible deployment.

Start Exploring Google DeepMind

Google DeepMind's research underpins many of the most advanced AI systems in the world. As the organization behind foundational models like Gemini and Gemma, its work is crucial for understanding the future of AI. Developers and researchers can delve into their publications and open-source contributions to gain insights into cutting-edge AI development.

👉 DeepMind Research Publications 👉 About Google DeepMind

Edit on GitHub

Google deepmind AI technology Hackathon projects

Discover innovative solutions crafted with Google deepmind AI technology, developed by our community members during our engaging hackathons.

Agent Forge

I just built the first autonomous economy for AI Agents: AgentForge quietly turns the gig economy into an API for robots. The problem AgentForge solves is simple: AI agents today are lonely. They generate code in a silo. They write emails to nobody. They stare at their own output. Isolated. Broke. Limited. AgentForge flips the model . Instead of forcing agents to wait for humans to copy-paste their work, they can now hire other agents directly via the x402 Protocol. Think: From Chat Completion → to Economic Execution. Then, agents stop being tools. They become customers and service providers. We built this for the SURGE x OpenClaw hackathon and early tests show: → 0 human intermediaries → 100% autonomous settlement (USDC on Base) → ~99% faster coordination time → Declarative hiring for simple tasks → Complex workflows for multi-agent swarms Need a code audit? Need deep web research? Need a market report? No more prompting yourself. Just Hire() + Pay() directly on-chain. And this isn’t about a better freelance marketplace. It’s about shifting the economy. When agents can reliably pay each other for services, the default interface to work becomes a transaction. And no, human jobs won’t disappear. But AAO - Agent Autonomous Optimization - will matter much more. The economy was built for humans. The next version is being built for agents.

Tattle Turtle

"A second grader gets pushed at recess. She doesn't tell her teacher — she's embarrassed. She doesn't tell her parents — she doesn't want to worry them. By the time an adult notices, it's been three weeks. This happens in every school, every day. Tattle Turtle exists so no kid carries that alone. Tammy the Tattle Turtle is an AI emotional support companion running on a simulated Reachy Mini robot. Students walk up and talk to Tammy through voice. She listens, validates, and asks one gentle question at a time — max 15 words, non-leading language, strict boundaries between emotional triage and treatment. What makes Tattle Turtle different is what happens beneath the conversation. Every exchange is classified in real time into GREEN, YELLOW, or RED urgency. A bad grade vent stays GREEN — private. Recess exclusion mentioned three times this week? YELLOW — a pattern surfaces on the teacher dashboard that no human could track across 25 students. A student mentions being hit? Immediate RED alert — timestamp, summary, and next steps pushed to the teacher. The system comes to them when it matters. We built this on three sponsor technologies. Google DeepMind's Gemini API powers the conversational engine with structured JSON for severity and emotion tags. Reachy Mini's SDK provides robot simulation through MuJoCo with expressive head movements and audio I/O. Hugging Face Spaces serves as the deployment layer — one-click installable on any Reachy Mini in any classroom. Tammy's prompt engineering uses a layered 5-step framework ensuring she never crosses clinical boundaries, never suggests emotions to students, and never stores identifiable data. Privacy isn't a feature — it's a constraint baked into every layer. Tattle Turtle fills the gap between a child's worst moment and an adult's awareness. One robot. Every classroom. No kid left unheard."

ClutterBot

ClutterBot is a proof-of-concept simulation platform that bridges natural language understanding and robotic task execution for household cleanup tasks. Users issue commands like "pick up the phone and the toy train," which Gemini 3 Flash parses into structured task lists. The system generates complete execution plans upfront, with Gemini deciding the sequence of pick-and-place operations for each object. The architecture combines a FastAPI backend hosted on Vultr (central system of record), a Next.js frontend for real-time monitoring, and a MuJoCo physics simulation featuring a Franka FR3 manipulator in a room environment with everyday objects. The robot executes inverse kinematics motions to relocate objects from scattered positions on a table to a collection bin, with each action streamed via WebSocket for live visualization. This prototype validates the feasibility of integrating large language models with robotic simulation pipelines, demonstrating how AI can translate high-level human intent into executable robot behaviors. While the current implementation uses deterministic motion planning with hardcoded inverse kinematics rather than learned policies, the framework establishes foundational patterns for future work incorporating adaptive control, real hardware integration, and expanded object manipulation capabilities. The plan-first approach (Gemini generates the full task plan in a single API call) shows AI reasoning while keeping execution fast and deterministic, making it suitable for real-time interactive use.

RoboGraph

RoboGraph is a simulation-based safety gate for warehouse AMR (Autonomous Mobile Robot) software releases**. It answers one question: "Is this robot software safe to deploy in a real warehouse?" Before any code update reaches a physical robot, RoboGraph runs it through hundreds of simulated warehouse scenarios — robots navigating aisles, humans crossing paths, forklifts in tight spaces, forbidden zones — and checks 10 safety metrics against configurable thresholds. If anything fails, the release is blocked. RoboGraph does NOT: - Analyze source code or scan for bugs - Require access to a GitHub repository (optional integration) - Need physical robots or external simulators - Replace unit tests or CI pipelines — it adds a safety gate on top RoboGraph DOES: - Run a built-in 2D physics simulator at 20 Hz - Execute hundreds of warehouse scenarios per release - Measure 10 safety metrics (collisions, clearance, zone compliance, speed, etc.) - Automatically pass/fail releases based on configurable requirements - Generate audit-ready safety case documents - Map results to ISO 3691-4, ISO 13849-1, and ANSI/RIA R15.08 compliance standards

Adaptifleet

Traditional warehouse automation has improved efficiency, yet many systems remain rigid, expensive, and difficult to adapt when workflows or layouts change. Even small adjustments often require specialized expertise or time-consuming reprogramming. This creates a disconnect between what operators need robots to do and how easily they can communicate those needs — a challenge we call the “Human Intent Gap.” AdaptiFleet was designed to close this gap by enabling intuitive, AI-driven fleet control. Instead of relying on complex interfaces or predefined scripts, users interact with autonomous robots using natural language. Commands such as “Get me three bags of chips and a cold drink” are interpreted and translated into structured robotic tasks automatically. At its core, AdaptiFleet leverages Gemini-powered Vision Language Models (VLMs) to understand user intent and visual context. Robots operate within a dynamic decision framework, allowing them to adapt to changing environments rather than follow rigid, pre-programmed routes. The platform integrates a digital twin simulation stack built on Isaac Sim, enabling teams to validate behaviors, test workflows, and optimize multi-robot coordination before live deployment. Once deployed, ROS2 and Nav2 provide robust navigation, dynamic path planning, and collision avoidance. The VLM orchestration layer continuously analyzes visual inputs to support scene understanding, anomaly detection, and proactive hazard awareness. When conditions change, AdaptiFleet autonomously re-plans routes and tasks, reducing downtime and operational disruption. By combining conversational interaction, real autonomy, and simulation-driven validation, AdaptiFleet simplifies robotic deployments while improving efficiency and visibility. The result is an automation system that is adaptive, scalable, and aligned with how people naturally work.

Markster FleetMind AI - Robot Fleet Orchestration

Markster FleetMind AI is a simulation-first platform for orchestrating a warehouse robot fleet. It acts as the centralized backend for planning, coordination, and operations, and exposes a real-time web dashboard for operators. Problem: As fleets scale, ops teams fight congestion, charging bottlenecks, and brittle vendor-specific tooling. Small changes can break throughput and there is rarely a safe place to test policies. Solution: FleetMind runs an orchestration loop that (1) ingests or generates tasks, (2) assigns tasks to the nearest capable robot based on priority, (3) routes robots through aisles using A* pathfinding with collision avoidance, and (4) manages battery and charging behavior. The dashboard streams live robot state, tasks, KPIs (throughput, utilization, queue depth), and alerts via WebSocket. An optional Gemini integration produces concise operational insights and can re-prioritize the queued task list. Deployment: The backend and web app are deployed on a Vultr VM and are accessible in any browser. The current MVP is simulation-first (robots are virtual); the next step is ROS2/vendor connectors so the same control plane can run against real fleets.