Heimdall

Heimdall is a cloud machine learning platform that takes a team from raw data to a live prediction API without a data science team. It combines a data warehouse, notebooks, no code model training, forecasting and one click deployment in one system, so data does not have to move between separate tools. It is aimed at teams that hold useful data but lack machine learning specialists. The platform is organized into five products. Lake is the warehouse: it ingests CSVs, database connections and unstructured zip archives and organizes them in a bronze, silver and gold catalog, where bronze holds raw uploads, silver holds cleaned and combined data, and gold holds datasets ready for modeling. Lab provides Python notebooks for profiling, exploration and transformation. ML trains classification and regression models without code. Forecast handles time series forecasting. Loop runs online learning models that keep adapting as new live data arrives. Trained models deploy as REST APIs in one click with monitoring built in, and the documentation covers API integration and database connections for production use. Example use cases on the site include real estate price prediction, ticket demand forecasting and personalized feed curation. Heimdall is free to start with a free account, but no plan names, tiers or prices are published on the site or in the documentation. Buyers should expect to learn paid terms after signing up. One point to note: production models and forecasts must be trained on gold layer data, so the catalog workflow is a requirement rather than an option.