ML Operations

Chalk

Real-time feature platform and inference runtime for production AI

About Chalk

Chalk provides a platform that supplies live context to models and agents through a feature store, real-time serving and federated queries, plus model serving, sandboxes and agent evals. It can be self-hosted in the customer's cloud or used as a serverless service.

Common use cases

  • Fraud and risk decisioning
  • Real-time personalization
  • Agent context
  • Model serving

Products and services

Context

Feature store, real-time serving and federated queries.

Runtime

Model serving, sandboxes, Model Gateway and functions.

Learning

Agent evals, notebooks and model training.

Deployment

Self-hosted and serverless options with observability.

Key features

  • Real-time features
  • Feature store
  • Model serving
  • Agent sandboxes
  • Federated queries
  • Self-hosted option

Integrations

  • Snowflake
  • Databricks
  • AWS
  • Kafka

Chalk alternatives

FAQs

Can it run in my cloud?

Yes, self-hosted in the customer's cloud or Chalk Serverless.

What does it do?

Provides features and context to models in real time and hosts inference.

Is engineering help offered?

The site lists forward-deployed engineers.

Work at Chalk? Request updates to this profile. Update this profile