Nick Kang, Product Manager on Kaggle Benchmarks, walks through how to build, run, and push an AI evaluation task in your own local development environment. He installs the write-kaggle-benchmarks from GitHub, initializes the environment, and builds a colorblind image test from scratch, asserting the correct output, validating it locally, and pushing it to Kaggle to run across four models: Gemini 3 Flash Preview, Gemini 3.5 Flash, GPT 5.5, and Opus 4.7. Results come back with pass/fail, latency, cost, and token counts. All visualized directly in the CLI and agent panel. 👉 Install the write-kaggle-benchmarks skill: https://github.com/Kaggle/kaggle-skills/tree/main/write-kaggle-benchmarks 👉 Get started on: https://www.kaggle.com/benchmarks About Kaggle: Kaggle’s global community of practitioners, researchers, and enthusiasts collaborate to shape the frontier of AI. Through AI competitions, benchmarks, agentic evaluation, Kaggle serves as both the engine and proving ground for community-led innovation. Follow Kaggle online 🌐 Visit the WEBSITE: https://www.kaggle.com ✍️ Check out our BLOG: https://www.kaggle.com/blog 🐦 Follow Kaggle on TWITTER: https://twitter.com/kaggle 🔗 Connect with us on LINKEDIN: https://www.linkedin.com/company/kaggle 🎮 Join us on the Kaggle Discord server: http://discord.gg/kaggle 🎥 Subscribe to our YouTube channel: https://www.youtube.com/@kaggle

Stay Updated with Kaggle Benchmark Notifications! We’ve introduced three new ways to track AI progress on Kaggle. Whether you're a benchmark owner or a follower, you can now get real-time alerts via web and email. About Kaggle: Kaggle’s global community of practitioners, researchers, and enthusiasts collaborate to shape the frontier of AI. Through AI competitions, benchmarks, agentic evaluation, Kaggle serves as both the engine and proving ground for community-led innovation. Follow Kaggle online 🌐 Visit the WEBSITE: https://www.kaggle.com ✍️ Check out our BLOG: https://www.kaggle.com/blog 🐦 Follow Kaggle on TWITTER: https://twitter.com/kaggle 🔗 Connect with us on LINKEDIN: https://www.linkedin.com/company/kaggle 🎮 Join us on the Kaggle Discord server: http://discord.gg/kaggle 🎥 Subscribe to our YouTube channel: https://www.youtube.com/@kaggle

Welcome to Kaggle's 5-Day AI Agents Intensive course with Google! This video is your essential, quick guide to navigating and using the Kaggle Notebooks environment for the intensive. What You'll Learn: Notebook Essentials - Whether you're new to Kaggle or need a quick refresher, this video covers all the basics you need to start coding effectively: - Understand notebooks as your interactive coding environment - Learn about code cells (for writing and executing code in Python) and markdown cells (for documentation) - Master how to edit cells, how to run cells, and the importance of the order of running cells in notebooks - See how to share notebooks with others for collaboration - How to add API keys via Kaggle Secrets so you don't save API keys as plaintext 🎮 Join us and other participants on Kaggle Discord server: http://discord.gg/kaggle

Read the whitepaper here: https://www.kaggle.com/whitepaper-introduction-to-agents Learn more about the 5-Day AI Agents Intensive: https://rsvp.withgoogle.com/events/google-ai-agents-intensive_2025 Introduction: Artificial intelligence is changing. For years, the focus has been on models that excel at passive, discrete tasks: answering a question, translating text, or generating an image from a prompt. This paradigm, while powerful, requires constant human direction for every step. We're now seeing a paradigm shift, moving from AI that just predicts or creates content to a new class of software capable of autonomous problem-solving and task execution. This new frontier is built around AI agents. An agent is not simply an AI model in a static workflow; it's a complete application, making plans and taking actions to achieve goals. It combines a Language Model's (LM) ability to reason with the practical ability to act, allowing it to handle complex, multi-step tasks that a model alone cannot. The critical capability is that agents can work on their own, figuring out the next steps needed to reach a goal without a person guiding them at every turn. 👉 Learn more about ADK here: https://google.github.io/adk-docs/ 🎮 Join us and other participants on Kaggle Discord server: http://discord.gg/kaggle

Read the whitepaper here: https://www.kaggle.com/whitepaper-agent-tools-and-interoperability-with-mcp Learn more about the 5-Day AI Agents Intensive: https://rsvp.withgoogle.com/events/google-ai-agents-intensive_2025 Introduction: In this whitepaper we talk first about the nature of tools used by foundation models: what they are and how to use them. We give some best practices and guidelines for designing effective tools and using them effectively. We then look at the Model Context Protocol, talking about its basic components and some of the challenges and risks it entails. Finally, we take a deeper look at the security challenges posed by MCP as it is introduced in an enterprise environment and connected to high-value external systems. 👉 Learn more about ADK here: https://google.github.io/adk-docs/ 🎮 Join us and other participants on Kaggle Discord server: http://discord.gg/kaggle

Read the whitepaper here: https://www.kaggle.com/whitepaper-context-engineering-sessions-and-memory Learn more about the 5-Day AI Agents Intensive: https://rsvp.withgoogle.com/events/google-ai-agents-intensive_2025 Introduction: This whitepaper explores the critical role of Sessions and Memory in building stateful, intelligent LLM agents to empower developers to create more powerful, personalized, and persistent AI experiences. To enable Large Language Models (LLMs) to remember, learn, and personalize interactions, developers must dynamically assemble and manage information within their context window—a process known as Context Engineering. 👉 Learn more about ADK here: https://google.github.io/adk-docs/ 🎮 Join us and other participants on Kaggle Discord server: http://discord.gg/kaggle

Read the whitepaper here: https://www.kaggle.com/whitepaper-agent-quality Learn more about the 5-Day AI Agents Intensive: https://rsvp.withgoogle.com/events/google-ai-agents-intensive_2025 Introduction: This is a practical guide from evaluation to observability. We are at the dawn of the agentic era. The transition from predictable, instruction-based tools to autonomous, goal-oriented AI agents presents one of the most profound shifts in software engineering in decades. While these agents unlock incredible capabilities, their inherent non-determinism makes them unpredictable and shatters our traditional models of quality assurance. This whitepaper serves as a practical guide to this new reality, founded on a simple but radical principle: agent quality is an architectural pillar, not a final testing phase. 👉 Learn more about ADK here: https://google.github.io/adk-docs/ 🎮 Join us and other participants on Kaggle Discord server: http://discord.gg/kaggle

Read the whitepaper here: https://www.kaggle.com/whitepaper-prototype-to-production Learn more about the 5-Day AI Agents Intensive: https://rsvp.withgoogle.com/events/google-ai-agents-intensive_2025 Introduction: This whitepaper provides a comprehensive technical guide to the operational life cycle of AI agents, focusing on deployment, scaling, and productionizing. Building on Day 4's coverage of evaluation and observability, this guide emphasizes how to build the necessary trust to move agents into production through robust CI/CD pipelines and scalable infrastructure. It explores the challenges of transitioning agent-based systems from prototypes to enterprise- grade solutions, with special attention to Agent2Agent (A2A) interoperability. This guide offers practical insights for AI/ML engineers, DevOps professionals, and system architects. 👉 Learn more about ADK here: https://google.github.io/adk-docs/ 🎮 Join us and other participants on Kaggle Discord server: http://discord.gg/kaggle

Join the Measuring Progress Toward AGI - Cognitive Abilities hackathon team for a technical breakdown of the "Measuring Progress Toward AGI: A Cognitive Framework" paper and its application to your hackathon tasks. Agenda: 20-Min Deep Dive: A walkthrough of the research paper, cognitive faculty examples, and exactly what the judges are looking for in a winning benchmark. 20-Min Live AMA: Your chance to ask the authors anything about the framework, task design, or scoring. The Panel: Ryan Burnell, Staff Research Scientist, Google DeepMind (https://x.com/DrRyanBurnell) Oran Kelly, Product Manager, Google DeepMind (https://www.linkedin.com/in/oran-kelly/) Nick Kang Product Manager, Kaggle Benchmarks (https://x.com/nick_kango) Enter the Hackathon: https://www.kaggle.com/competitions/kaggle-measuring-agi The Paper: https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/measuring-progress-toward-agi/measuring-progress-toward-agi-a-cognitive-framework.pdf SUBSCRIBE: https://www.youtube.com/c/kaggle?sub_confirmation=1&utm_medium=youtube&utm_source=channel&utm_campaign=yt-sub About Kaggle: Kaggle’s global community of practitioners, researchers, and enthusiasts collaborate to shape the frontier of AI. Through AI competitions, benchmarks, agentic evaluation, Kaggle serves as both the engine and proving ground for community-led innovation. Follow Kaggle online: Visit the WEBSITE: http://www.kaggle.com/?utm_medium=youtube&utm_source=channel&utm_campaign=yt-kg Like Kaggle on FACEBOOK: http://www.facebook.com/kaggle?utm_medium=youtube&utm_source=channel&utm_campaign=yt-fb Follow Kaggle on TWITTER: http://twitter.com/kaggle?utm_medium=youtube&utm_source=channel&utm_campaign=yt-tw Check out our BLOG: http://blog.kaggle.com/?utm_medium=youtube&utm_source=channel&utm_campaign=yt-blog Connect with us on LINKEDIN: http://www.linkedin.com/company/kaggle?utm_medium=youtube&utm_source=channel&utm_campaign=yt-lkn Advance your data science skills: Take our free online courses: http://www.kaggle.com/learn/overview?utm_medium=youtube&utm_source=channel&utm_campaign=yt-learn Get started with Kaggle Kernels: http://www.kaggle.com/docs/kernels?utm_medium=youtube&utm_source=channel&utm_campaign=yt-krnl Download clean datasets from Kaggle: http://www.kaggle.com/docs/datasets?utm_medium=youtube&utm_source=channel&utm_campaign=yt-datast Sign up for a Kaggle Competition: http://www.kaggle.com/docs/competitions?utm_medium=youtube&utm_source=channel&utm_campaign=yt-comps Explore the Kaggle Public API: http://www.kaggle.com/docs/api?utm_medium=youtube&utm_source=channel&utm_campaign=yt-docs Kaggle https://www.youtube.com/c/kaggle

Join the Measuring Progress Toward AGI - Cognitive Abilities hackathon team for a technical breakdown of the "Measuring Progress Toward AGI: A Cognitive Framework" paper and its application to your hackathon tasks. Agenda: 20-Min Deep Dive: A walkthrough of the research paper, cognitive faculty examples, and exactly what the judges are looking for in a winning benchmark. 20-Min Live AMA: Your chance to ask the authors anything about the framework, task design, or scoring. The Panel: Ryan Burnell, Staff Research Scientist, Google DeepMind (https://x.com/DrRyanBurnell) Oran Kelly, Product Manager, Google DeepMind (https://www.linkedin.com/in/oran-kelly/) Nick Kang Product Manager, Kaggle Benchmarks (https://x.com/nick_kango) Enter the Hackathon: https://www.kaggle.com/competitions/kaggle-measuring-agi The Paper: https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/measuring-progress-toward-agi/measuring-progress-toward-agi-a-cognitive-framework.pdf SUBSCRIBE: https://www.youtube.com/c/kaggle?sub_confirmation=1&utm_medium=youtube&utm_source=channel&utm_campaign=yt-sub About Kaggle: Kaggle is the world's largest community of data scientists. Join us to compete, collaborate, learn, and do your data science work. Kaggle's platform is the fastest way to get started on a new data science project. Spin up a Jupyter notebook with a single click. Build with our huge repository of free code and data. Stumped? Ask the friendly Kaggle community for help. Follow Kaggle online: Visit the WEBSITE: http://www.kaggle.com/?utm_medium=youtube&utm_source=channel&utm_campaign=yt-kg Like Kaggle on FACEBOOK: http://www.facebook.com/kaggle?utm_medium=youtube&utm_source=channel&utm_campaign=yt-fb Follow Kaggle on TWITTER: http://twitter.com/kaggle?utm_medium=youtube&utm_source=channel&utm_campaign=yt-tw Check out our BLOG: http://blog.kaggle.com/?utm_medium=youtube&utm_source=channel&utm_campaign=yt-blog Connect with us on LINKEDIN: http://www.linkedin.com/company/kaggle?utm_medium=youtube&utm_source=channel&utm_campaign=yt-lkn Advance your data science skills: Take our free online courses: http://www.kaggle.com/learn/overview?utm_medium=youtube&utm_source=channel&utm_campaign=yt-learn Get started with Kaggle Kernels: http://www.kaggle.com/docs/kernels?utm_medium=youtube&utm_source=channel&utm_campaign=yt-krnl Download clean datasets from Kaggle: http://www.kaggle.com/docs/datasets?utm_medium=youtube&utm_source=channel&utm_campaign=yt-datast Sign up for a Kaggle Competition: http://www.kaggle.com/docs/competitions?utm_medium=youtube&utm_source=channel&utm_campaign=yt-comps Explore the Kaggle Public API: http://www.kaggle.com/docs/api?utm_medium=youtube&utm_source=channel&utm_campaign=yt-docs Kaggle https://www.youtube.com/c/kaggle

Learn about the winning approach by Team Patrick Yam and the key techniques they used in Jane Street Real-Time Market Data Forecasting competition. This competition hosted by Jane Street, challenged participants to use real-world data from production trading systems to build a model that can predict and navigate the difficult challenges of financial markets, including fat-tailed distributions, non-stationary time series, and sudden market shifts 👉 Competition: https://www.kaggle.com/competitions/jane-street-real-time-market-data-forecasting/overview 🏆 Leaderboard: https://www.kaggle.com/competitions/jane-street-real-time-market-data-forecasting/leaderboard About Kaggle: Kaggle is the world's largest community of data scientists. Join us to compete, collaborate, learn, and do your data science work. Kaggle's platform is the fastest way to get started on a new data science project. Spin up a Jupyter notebook with a single click. Build with our huge repository of free code and data. Stumped? Ask the friendly Kaggle community for help. Follow Kaggle online 🌐 Visit the WEBSITE: https://www.kaggle.com ✍️ Check out our BLOG: https://www.kaggle.com/blog 🐦 Follow Kaggle on TWITTER: https://twitter.com/kaggle 🔗 Connect with us on LINKEDIN: https://www.linkedin.com/company/kaggle 🎮 Join us on the Kaggle Discord server: http://discord.gg/kaggle 🎥 Subscribe to our YouTube channel: https://www.youtube.com/@kaggle #Kaggle #DataScience #QuantitativeTrading #Finance #MachineLearning

Learn about the winning approach by TeamZ Lab 数据实验室 and the key techniques they used in LEAP - Atmospheric Physics using AI (ClimSim) competition. This competition tasks participants with developing machine learning emulators to represent subgrid-scale atmospheric processes, such as storms and turbulence, within operational climate models. By approximating complex physical interactions at a fraction of the computational cost of traditional simulations, the challenge seeks to reduce uncertainty in global warming projections and provide policymakers with more accessible, high-resolution climate data. 👉 Competition: https://www.kaggle.com/competitions/leap-atmospheric-physics-ai-climsim/overview 🏆 Leaderboard: https://www.kaggle.com/competitions/leap-atmospheric-physics-ai-climsim/leaderboard About Kaggle: Kaggle is the world's largest community of data scientists. Join us to compete, collaborate, learn, and do your data science work. Kaggle's platform is the fastest way to get started on a new data science project. Spin up a Jupyter notebook with a single click. Build with our huge repository of free code and data. Stumped? Ask the friendly Kaggle community for help. Follow Kaggle online 🌐 Visit the WEBSITE: https://www.kaggle.com ✍️ Check out our BLOG: https://www.kaggle.com/blog 🐦 Follow Kaggle on TWITTER: https://twitter.com/kaggle 🔗 Connect with us on LINKEDIN: https://www.linkedin.com/company/kaggle 🎮 Join us on the Kaggle Discord server: http://discord.gg/kaggle 🎥 Subscribe to our YouTube channel: https://www.youtube.com/@kaggle #Kaggle #DataScience #MachineLearning

Learn about the winning approach by GreySnow and the key techniques they used in LEAP - Atmospheric Physics using AI (ClimSim) competition. This competition tasks participants with developing machine learning emulators to represent subgrid-scale atmospheric processes, such as storms and turbulence, within operational climate models. By approximating complex physical interactions at a fraction of the computational cost of traditional simulations, the challenge seeks to reduce uncertainty in global warming projections and provide policymakers with more accessible, high-resolution climate data. 👉 Competition: https://www.kaggle.com/competitions/leap-atmospheric-physics-ai-climsim/overview 🏆 Leaderboard: https://www.kaggle.com/competitions/leap-atmospheric-physics-ai-climsim/leaderboard About Kaggle: Kaggle is the world's largest community of data scientists. Join us to compete, collaborate, learn, and do your data science work. Kaggle's platform is the fastest way to get started on a new data science project. Spin up a Jupyter notebook with a single click. Build with our huge repository of free code and data. Stumped? Ask the friendly Kaggle community for help. Follow Kaggle online 🌐 Visit the WEBSITE: https://www.kaggle.com ✍️ Check out our BLOG: https://www.kaggle.com/blog 🐦 Follow Kaggle on TWITTER: https://twitter.com/kaggle 🔗 Connect with us on LINKEDIN: https://www.linkedin.com/company/kaggle 🎮 Join us on the Kaggle Discord server: http://discord.gg/kaggle 🎥 Subscribe to our YouTube channel: https://www.youtube.com/@kaggle #Kaggle #DataScience #MachineLearning

Learn about the winning approach by Team 预测多了一点 and the key techniques they used in Enefit - Predict Energy Behavior of Prosumers competition. This competition tasks participants with developing an energy prediction model to accurately forecast the power generation and consumption of prosumers within the Baltic energy grid. By minimizing the discrepancy between expected and actual energy use, the challenge seeks to reduce high imbalance costs for providers like Enefit and improve grid stability for a more sustainable renewable energy infrastructure. 👉 Competition: https://www.kaggle.com/competitions/predict-energy-behavior-of-prosumers/overview 🏆 Leaderboard: https://www.kaggle.com/competitions/predict-energy-behavior-of-prosumers/leaderboard About Kaggle: Kaggle is the world's largest community of data scientists. Join us to compete, collaborate, learn, and do your data science work. Kaggle's platform is the fastest way to get started on a new data science project. Spin up a Jupyter notebook with a single click. Build with our huge repository of free code and data. Stumped? Ask the friendly Kaggle community for help. Follow Kaggle online 🌐 Visit the WEBSITE: https://www.kaggle.com ✍️ Check out our BLOG: https://www.kaggle.com/blog 🐦 Follow Kaggle on TWITTER: https://twitter.com/kaggle 🔗 Connect with us on LINKEDIN: https://www.linkedin.com/company/kaggle 🎮 Join us on the Kaggle Discord server: http://discord.gg/kaggle 🎥 Subscribe to our YouTube channel: https://www.youtube.com/@kaggle #Kaggle #DataScience #MachineLearning

Learn about the winning approach by HYD and the key techniques they used in Enefit - Predict Energy Behavior of Prosumers competition. This competition tasks participants with developing an energy prediction model to accurately forecast the power generation and consumption of prosumers within the Baltic energy grid. By minimizing the discrepancy between expected and actual energy use, the challenge seeks to reduce high imbalance costs for providers like Enefit and improve grid stability for a more sustainable renewable energy infrastructure. 👉 Competition: https://www.kaggle.com/competitions/predict-energy-behavior-of-prosumers/overview 🏆 Leaderboard: https://www.kaggle.com/competitions/predict-energy-behavior-of-prosumers/leaderboard About Kaggle: Kaggle is the world's largest community of data scientists. Join us to compete, collaborate, learn, and do your data science work. Kaggle's platform is the fastest way to get started on a new data science project. Spin up a Jupyter notebook with a single click. Build with our huge repository of free code and data. Stumped? Ask the friendly Kaggle community for help. Follow Kaggle online 🌐 Visit the WEBSITE: https://www.kaggle.com ✍️ Check out our BLOG: https://www.kaggle.com/blog 🐦 Follow Kaggle on TWITTER: https://twitter.com/kaggle 🔗 Connect with us on LINKEDIN: https://www.linkedin.com/company/kaggle 🎮 Join us on the Kaggle Discord server: http://discord.gg/kaggle 🎥 Subscribe to our YouTube channel: https://www.youtube.com/@kaggle #Kaggle #DataScience #MachineLearning