Chalk
Feature platform for real-time machine learning data
Chalk is profiled here as a Vector Database tool for engineering teams. Read about features, pricing, and how it compares to related options in the tools directory.
Description
Chalk is a data platform for machine learning founded in 2021 that computes features in real time from a declarative Python definition. Engineers describe features and their data sources as code, and Chalk builds a computation graph that fetches and transforms data on demand at inference time, with caching and freshness controls per feature. Fraud, underwriting, and recommendation systems use it where features must reflect data as of the moment of a request. Computing features at request time keeps inputs current, which matters for fraud, underwriting, and recommendation systems making live decisions. Per-feature caching and freshness controls let teams tune cost against how recent each value must be.
Key Capabilities:
Features defined declaratively in Python with typed resolvers
On-demand real-time feature computation at inference time
Online and offline stores with point-in-time correctness
Per-feature caching and freshness configuration
Streaming and batch source integration
Deployment in the customer's own cloud environment