DBT in ML

Integrate ML Models with your DBT DAG in SQL.

Use ML pipelines, make them part of your DAG, and scheduled jobs without leaving DBT.

ML with SQL

Train ML models with your dbt models and share them with your team. No matter which data warehouse you use, you can train AutoML models and make predictions with SQL in dbt. It's easy with Layer metadata store.

SELECT
 id,
 layer.predict(
'layer/models/review_score_predictor',
 review)

FROM
 {{ ref ('products') }]

Train AutoML

Layer enables you train AutoML models in your dbt DAG with SQL. When you register your models (with simply an @model decorator) and log your metrics, all your metadata is seamlessly organized in a single place.

SELECT
 order_id,
 layer.automl(
'regressor',
 ARRAY[
 days_between_purchase_and_delivery,
 order_approved_late,
 actual_delivery_vs_expectation,
 total_order_price,   
 total_order_freight,
 is_multiItems_order,
 seller_shipped_late],
 review_score
)

FROM
 {{ ref ('training_data') }]

Predict

Layer enables you to score your data or train AutoML models inside dbt DAG. Using your dbt models, train a churn model to predict churning users or train a product recommendation models for your marketing team.

SELECT
 id,
 layer.predict(
'layer/clothing/models/objectdetection',
 ARRAY[image])

FROM
 {{ ref ('products') }]

Share

Layer solves the headache of inaccessible data, unreproducible experiments and scattered documentation that results in harsh project handovers. In Layer, each project has a dedicated spaces where all documentation and ML entities are found in one place allowing easy presentation and communication.

Community

The definitive ML project development diary

Show off your work with demo-able & easy-to-present dynamic content for your ML projects.

See community examples
Corporate

Built with data teams, for data teams

Layer is the single place for all datasets, models and documents developed by your team.

Request a demo