Accelerate AI success with Dataiku#
Dataiku brings together five core capabilities for agentic AI, machine learning, modern analytics, orchestration, and governance in one unified enterprise platform.
Agentic transformation: Design, deploy, and improve AI agents powered by enterprise data, analytics, and business logic, with decision governance built in, not bolted on.
Visual Agents & Code Agents: Build structured, multi-step agents using visual logic blocks or full-code frameworks. Combine prompts, tools, APIs, and models into governed reasoning workflows.
Users can read the AI Agents section (in the reference documentation).
Coders will find multiple tutorials in Agents and Tools for Generative AI, as well as code samples on the Agents page.
For more information, non-coders can follow the Agents section (in the Knowledge Base).
Agent Hub & multi-agent orchestration: Centralize agent creation, collaboration, lifecycle management, and orchestration. Route tasks across agents, models, and tools in a shared control plane.
Users can read the Agent Hub section (in the reference documentation).
Coders and non-coders will find a lot of information on their respective websites (the Developer Guide and the Knowledge Base, respectively).
The Dataiku LLM Mesh: Abstract and switch LLM providers with centralized routing, quotas, monitoring, and safety controls to manage cost and risk at scale.
Users can read the Generative AI and LLM Mesh section (in the reference documentation).
Coders will find multiple tutorials in Generative AI, as well as code samples on the LLM Mesh page.
For more information, non-coders can follow the LLM Administration section (in the Knowledge Base).
Scalable Machine Learning: Unify visual ML, AutoML, and full-code development in one governed environment, so models move from experimentation to production without silos.
Visual ML & AutoML: Enable analysts and data scientists to build, validate, and compare models using code-free and assisted workflows while preserving flexibility.
Users can read the Machine learning section (in the reference documentation).
Coders will find multiple tutorials in Machine Learning, as well as code samples on the Visual Machine learning page.
For more information, non-coders can follow the Leverage Machine Learning (in the Knowledge Base).
Unified Flow & deployment automation: Move from experimentation to APIs, batch scoring, or production pipelines without handoffs. Package, deploy, and monitor models within one system.
Users can read the MLOps section (in the reference documentation).
Coders will find multiple tutorials in xOps and code samples in several pages.
For more information, non-coders can follow the AIOps Overview (in the Knowledge Base).
Model registry & lifecycle governance: Centralize model registries, approval workflows, lineage, performance monitoring, and regulatory documentation within a single, governed framework.
Users can read the AI Governance section (in the reference documentation).
Coders will find multiple tutorials on AI Governance and code samples on the Dataiku Govern page.
For more information, non-coders can follow the Leverage Machine Learning (in the Knowledge Base).
Modernized Analytics: Standardize data preparation, accelerate workflows with AI assistance, and deliver trusted analytics, while working directly on your existing platforms.
Visual & code recipes: Join, clean, and transform data using visual recipes or Python, R, and SQL, fully documented and versioned in the Flow.
Users can read the Data preparation section (in the reference documentation).
Coders will find multiple tutorials on Data Engineering and code samples on the Recipes page.
For more information, non-coders can follow the Prepare and Transform Data (in the Knowledge Base).
AI assistants: Generate preparation steps, SQL queries, and Flow creation plans using natural language, with full transparency and governance.
Users can read the AI Assistants section (in the reference documentation).
Coders will find multiple tutorials in Code assistants in Code Studios.
For more information, non-coders will find a lot of information on the AI Assistants in the Knowledge Base.
Data lineage, quality rules, & certification: Embed lineage tracking, validation checks, documentation, and dataset certification directly into analytics pipelines.
Users can read the Metrics, checks and Data Quality section (in the reference documentation).
Coders will find multiple code samples in Data Quality.
For more information, non-coders can follow the Explore Data Quality (in the Knowledge Base).
AI Orchestration: Orchestrate agents, analytics, models, data, and tools into a unified execution layer, so AI runs as part of the business, not beside it.
Extensive data connectors & native compute: Connect to Snowflake, Databricks, Redshift, S3, APIs, and enterprise systems while leveraging SQL, Spark, Kubernetes, or GPUs.
Users can read the Connecting to data section (in the reference documentation).
Coders can learn how to create and use a custom connector on the Plugins development page.
For more information, non-coders can follow the Import Data (in the Knowledge Base).
Unified execution layer: Run agents, analytics workflows, and ML models in shared production pipelines with built-in automation and monitoring.
Users can read the Automation scenarios section (in the reference documentation).
Coders will find multiple code samples in the “”Concepts and examples”” section.
For more information, non-coders can follow the Operating Dataiku (in the Knowledge Base).
Third-party agent orchestration: Manage agents across Snowflake Cortex, Databricks, Vertex AI, Bedrock, and more, in a single governed system.
Coders and non-coders will find a lot of information across their respective websites.
AI Governance: Define policies, approvals, lineage, monitoring, and risk controls across analytics, models, and agents — inside Dataiku and beyond it.
Centralized AI inventory: Maintain a unified registry of datasets, models, agents, and projects with ownership, status, and approval tracking.
End-to-end lineage & explainability: Trace how data, features, models, and agent decisions connect, with built-in documentation and model explainability tools.
Continuous monitoring & risk controls: Monitor model drift, bias, LLM usage, cost, and performance across the lifecycle with structured approval workflows.
Coders and non-coders will find extensive coverage of these topics across their respective documentation, as these foundational Dataiku concepts are explored throughout multiple pages and use cases.