Machine Learning
Train, evaluate, and deploy models with conversational AutoML
Point ML Clever at your dataset, ask for algorithm comparisons, request feature importance breakdowns, and deploy the winning model with a follow-up prompt.
Automation across the pipeline
Let AutoML handle the heavy lifting
Choose your target column and let the platform profile data, handle missing values, scale features, and encode categories.
Run XGBoost, LightGBM, RandomForest, and more in parallel while the AI ranks results against your business metric.
Once the best family wins, AutoML fine-tunes configurations to squeeze out extra lift without manual search.
Guided setup
Configure training in seconds
Pick quick, balanced, or comprehensive modes, define your target, and set cost guards. The guided wizard previews transformations before you commit.
Operational efficiency
Better models, faster delivery
Leaderboards stack algorithms against your goal metric, with confidence intervals and calibration charts.
Generate feature importance, SHAP plots, and natural-language rationale ready for stakeholders.
Ship the champion model to prediction endpoints, scheduled batches, and interactive what-if apps with one click.
Built to work with the suite
A fully integrated ML workflow
AutoML works seamlessly with dashboards, projects, and presentations so insights stay coherent.
Transparent results
Understand every model
Compare models with accuracy, precision, recall, AUC, RMSE, or cost-weighted metrics. Leaderboards include confusion matrices, calibration plots, and narrative takeaways so non-scientists trust the outcome.
Platform workflows
Connect AutoML to production
Move from raw datasets to deployed prediction services without leaving ML Clever.
AutoML Prompt Library
Prompts and follow-ups for smarter runs
Prompt: Train a classification model to predict customer churn.
Follow-up: Show SHAP values for the top five features and recommend retention actions.
Prompt: Build a regression model forecasting monthly revenue.
Follow-up: Compare linear models with gradient boosting and share confidence intervals.
Prompt: Set up automated preprocessing for a mixed-type dataset.
Follow-up: Let me preview transformations and override encoding choices before training.
Prompt: Benchmark new models against last quarter's champion.
Follow-up: Highlight performance deltas and recommend whether to promote or keep the current model.
Prompt: Deploy the winning model to the prediction API.
Follow-up: Schedule nightly batch scoring and send results to the finance project.
Prompt: Set up monitoring and drift alerts.
Follow-up: Notify the data science team when accuracy drops below 2% of baseline.
Industries & roles
Where AutoML accelerates outcomes
Apply automated model training to high-impact use cases.
Resources
Master automated machine learning
Get tactical advice on bringing AutoML into your delivery model.
Train and deploy your next model in one session
Upload a dataset, let AutoML test and tune algorithms, and push the winning model live with built-in dashboards and prediction endpoints.
Each run includes performance leaderboards, explainability packs, and one-click deployment into production.
Frequently Asked Questions
Have questions? We have answers. If you can‘t find what you‘re looking for, feel free to contact us.
Select quick, balanced, or comprehensive modes and the system profiles your dataset automatically.
Can I review feature transformations before the run starts?
Yes. Preview planned preprocessing, adjust encodings, and lock business rules before training begins.
Use the algorithm panel to whitelist or blacklist learners, set search budgets, and define custom parameter grids.
Can I include or exclude models like XGBoost, LightGBM, or neural nets?
The AI still recommends best-fit options based on your data shape and goal metric.
Pick the primary metric and optionally supply cost matrices or weights so the leaderboard reflects your business reality.
Can I optimize for cost-sensitive metrics or business-weighted objectives?
Calibration charts, confusion matrices, and narrative summaries explain how each model performs.
Every run delivers explainability packs with feature importance, SHAP, partial dependence, and narrative breakdowns.
Does the AI provide feature importance, SHAP plots, and natural-language rationale?
Embed these insights directly into dashboards or presentations for stakeholders.
One click promotes the model to REST endpoints, scheduled batch jobs, and guided prediction apps.
Can I publish APIs, batch jobs, and interactive what-if tools simultaneously?
Business teams and engineers share the same scoring logic without rebuilding pipelines.
Monitoring dashboards track accuracy, latency, drift, and input health with configurable thresholds.
Will I get alerts when accuracy drops or data shifts?
Alerts can trigger retraining suggestions, notifications, or automated rollback.
Model outputs sync to Projects, feed dashboards with performance summaries, and refresh presentation slides with new prediction highlights.
Can dashboards and decks pull the latest metrics and explanations?
Stakeholders always see the latest metrics without manual exports.
Set runtime and cost limits, pin runs to specific regions, and download full experiment logs for governance reviews.
Can I cap runtime, specify regions, or export audit logs?
Audit trails include data lineage, configuration, and deployment events for regulated teams.