Applications

  • MODEL BUILDING & TRAINING
  • INTEGRATION TESTING & INCIDENT RESPONSE
  • DEMOS & POCS
  • DATA SHARING WITH REMOTE TEAMS

Drivers

DATA SPARSITY
  • Lack of sufficient data points
  • Class imbalance and the need for conditional generation and amplification of signals
DATA SHARING CONSTRAINTS
  • Privacy/confidentiality
  • Regulatory/compliance
  • Date leakage/lineage

Data Scientists

  • Need a Synthetic Data Workbench to meet the needs of various stakeholders
  • Adapts to any type of data
  • Produces high-fidelity & privacy-compliant synthetic data that augments and optimizes for downstream use cases

The Rockfish Generative AI Solution

Works with a variety of high-value enterprise data types for any industry vertical
  • Time series
  • Session journey
  • Tabular
  • Other
Fits seamlessly into existing Data and MLOps pipelines
  • Works as a shadow pipeline
  • Maintain the same data schema
Flexible deployment options
  • SaaS
  • Enterprise VPC (cloud agnostic)

Open Source

The Rockfish solution draws upon foundational innovation originally developed at Carnegie Mellon University that are open sourced
DataFuel Open Source Landing Repository link

Related Papers

  1. Generating Time Series Using GANs link
  2. Generating rare samples using GANs link
  3. Theoretical foundations on stability and privacy link , link
  4. Generating high fidelity network packet and flow traces link​​

Rockfish Data Overview Video

For a demo or for more information, please