SCRIBBLE

ENRICH

Feature Engineering, Data Pipelines
& Nothing More.
What is Scribble Enrich?

Enrich is a customizable Feature Store, built for data science teams that need efficiency and trust in their datasets.

 

Enrich ensures that each feature, and so each dataset built using these features, is reproducible, versioned, quality-checked, and searchable.

This means data science teams can deploy models that much faster, and with that much more confidence in the underlying data. This means faster re-training or debugging, and quicker turnaround time for each new version of these models.

 

 

Enrich is:

  • Highly customizable, and gives ML engineers an SDK

  • Streamlines data transforms and manages complexity using versioned pipelines to ease retraining and debugging

  • Eases collaboration across data teams via its feature marketplace

  • Provide target data sets feeding APIs for applications

  • Helps address emerging requirements of explainability, provenance, auditability.

CAPABILITIES

& BENEFITS

Data Transforms

Modules with parameter validation, input/ output validation, documentation

01

Feature Marketplace

App to discover datasets computed with any statistical attributes.

04

Metadata Tracking

Data accesses, and writes, process, quality metrics, state changes

02

Built-in Apps

Multiple apps including lightweight catalog, lineage search, simple labeller etc.

05

Controlled Deployment

Deploy from Github and online upgrade

03

Meta-Computation

Compute over data and metadata such as lineage, drift, expectations etc.

06

SCRIBBLE

ENRICH

Components & Architecture

Usecases

Track utilization of the features along with ownership

Implement

SDK and other services to rapidly implement feature engineering modules

Operate

Administer versioned, auditable, parameterized pipelines, each generating multiple data sets.

Audit

Check provenance of datasets by name or other attributes, and compare runs

Access

Discover datasets via a marketplace for features and along with search interface to build cohorts for analysis

Monitor

Check drift and access other custom usage monitoring services

How it Works

Enrich handles the complexity of computation and data semantics by providing a python SDK to develop, document and test the feature engineering modules (transforms, pipelines, scheduling, etc) and controlled execution on the server-side. 

 

The server provides an interface to discover, operate and audit the resulting features or datasets.

Hooks at either end of Enrich allow for understanding (cataloguing) input data stores, and surfacing features at any frequency through APIs for downstream consumption, by defining data contracts and integration points. 

 

So for Data Scientists, the Enrich feature store experience simplifies, standardizes, and speeds up the model development process, with confidence in their performance.

PRINCIPLED DESIGN

to Enable 10x Use Cases

hello@scribbledata.io

+1-720-445-8387

P-47, 2nd Floor, 6th Main, 8th Cross, LIC Colony, Jeevan Bima Nagar, Bangalore 560075

Indiqube Orion:

24th Main Rd, Garden

Layout, HSR Layout, 

Bengaluru 560102 

2035, Sunset Lake Road, Suite B-2,

Newark, DE - 19702

CIN: U72900KA2016PTC097462

  • White LinkedIn Icon
  • White Facebook Icon
  • White Twitter Icon

© 2020 by ScribbleData.