You know a Machine Learning solution is right for you, but could use help selecting a model or building one yourself.

POWERUP Machine Learning Minimum Viable Model helps you through the iterative process of developing an initial model based on your specific business use case. You’ll end up with a custom solution and a better understanding of where to take it next.

Or perhaps your existing ML model performs well, but you could take advantage of a pipeline built in Google Cloud. POWERUP Machine Learning Minimum Viable Pipeline provides a scalable, composable, and secure ML pipeline in Google Cloud to easily and quickly bring your ML solutions to market.

POWERUP Machine Learning Minimum Viable Model and/or Pipeline proceeds through five milestones: 

Minimum Viable Model 
Activities:

  • Milestone 1:  Explore data
    Perform analysis to understand your data
  • Milestone 2: Select algorithm
    Identify data science and machine learning approach
  • Milestone 3:  Develop feature engineering pipeline
    Create the right features from raw data for the ML task
  • Milestone 4: Build ML model 
    Develop, train, and evaluate the model on test datasets
  • Milestone 5: Iterate on ML model
    Refine ML model to improve performance and efficacy
  • Milestone 6: Finalize documentation
    Document and hand-off the ML model

Deliverables for the Minimum Viable Model

  • An initial ML model including code to train your model and evaluate its quality
  • A technical design document describing the model's architecture with recommended next steps on model iteration

Minimum Viable Pipeline 
Activities:

  • Milestone 1: Gather requirements
    Review infrastructure and gather requirements for pipeline needs
  • Milestone 2: Understand model 
    Understand the business goal and data required for the model
  • Milestone 3: Identify composable components of the pipeline
    Confirm that your design meets operational requirements
  • Milestone 4: Create ML Pipeline
    Generate pipeline in your Google Cloud account
  • Milestone 5: Document work and onboard users
    Perform testing to confirm users' ability to understand and run the pipeline
  • Milestone 6: Finalize documentation
    Write and deliver technical design document

Deliverables for the Minimum Viable Pipeline

  • A secure ML pipeline deployed on Google Cloud to quickly bring your ML solution to market
  • A technical design document with an inventory of Google Cloud configuration decisions
  • Pipeline code and scripts to launch pipelines 

Contact us to bring your ML proof-of-concept to market!