Data Intelligence product
Train classification and regression
Without writing training code
Train classification or regression on gold datasets, review why the model decided, then deploy a REST API — no training code.
Start with a CSV or a database. No data science team required.
Your Models at a Glance
Real-time insights into your machine learning models' performance
2 hours ago
Training Data
Customer records
45,230
samples
1 day ago
Training Data
Sales history
36,720
samples
3 hours ago
Training Data
Transaction records
156K
samples
Connect Your Data Sources
Effortlessly import your data from CSV files, Databricks, Oracle, and more. Train directly on Lake gold datasets or upload files — with more integrations on the roadmap.
File Upload
CSV, XLSX
Lake
Gold datasets
Databricks
Analytics
PostgreSQL
Database
MySQL
Database
MariaDB
Database
Understand Your Models
At Heimdall, we value model explainability and give you the necessary tools to understand what factors impact the performance and accuracy of your models.
Use cases
ML use cases on Heimdall
Classification and regression deployments teams run on gold datasets — from real estate pricing to operational risk.
See warehouse workflows that feed these models on all Heimdall use cases.
ML FAQ
- What is Heimdall ML?
- Heimdall ML is a no-code machine learning product for classification and regression. Teams train explainable models on Lake gold datasets or uploaded files, review feature importance, and deploy to production REST APIs with built-in usage monitoring.
- How does Heimdall ML compare to Amazon SageMaker?
- SageMaker is a developer-centric ML platform requiring notebooks, pipelines, and infrastructure configuration. Heimdall ML is no-code: select a gold dataset, configure targets and features, train, and deploy to a REST API in minutes — designed for analysts and product teams, not ML engineers managing endpoints.
- Does Heimdall ML require coding?
- No. Heimdall ML trains and deploys models through a visual interface. Teams can optionally use Lab (Python notebooks) upstream to prepare silver datasets, but model training and API deployment do not require writing code.
- What model types does Heimdall ML support?
- Heimdall ML supports classification (e.g., churn, fraud detection) and regression (e.g., price prediction, demand estimation) with explainable feature importance on every trained model.
Train your first model on gold data
Upload or connect data in Lake, publish gold, then deploy a classification or regression API you can call from your app.