WHAT WE OFFER
Data Science Assessment:
- Inspect and understand data environment and ability to provide data to build/deploy a pipeline.
- Consider internal talent and tech stack with ability to meet data science needs.
- Evaluate past projects and provide feedback on transparency & reproducability of models.
Custom Model Development:
- Build solution as a pipeline.
- Build and evaluate relevant candidate model solutions.
- Employ Databricks to help companies visualize data through dashboards and analyze data using machine learning (MLib), GraphX and Spark SQL.
- Deliver model development codebase.
- Can transform development codebase into deployment pipeline to fit into your tech ecosystem.
- Document model development process to ensure reproducability and transparency.
- Validate delivered model.
Model Development Infrastructure as a Service (MDaaS):
- Provide hosted model development environment for your data science staff to work in.
- Scalable to accommodate data storage and computational demands.
- Cloud-based, taking advantage of industry standard open source tech stack.
- Secure, git-based source code repository.
- Persist development data as necessary; valuable for on-going model validation.
Model Execution as a Service (MaaS):
- Hosted solution to score records in real-time or in batch.
- Can execute models developed by your Data Science team or by DecisivEdge.
- Integrate machine learning pipeline into your workflow via API call.
- Provide model execution monitoring.
- Model execution data retained for model performance tracking.
3rd Party Model Validation:
- DecisivEdge serves as independent 3rd party model validator.
- Utilize your model validation template to address specific regulatory environment or internal risk requirements.
- Outline model Risks, Findings, and Observations.
- Ongoing stability and validity performance monitoring.
- For regulated financial institutions, will deliver a Validation to satisfy OCC 2011-2012 model risk management guidelines.
- DecisivEdge help financial service providers with proper implementation of CECL according to their new accounting standards.
Data Science as a Service (DSaaS):
- Address resourcing needs in two capacities:
- Dedicated hours over a specific time period to direct as needed.
- Provide fractional resources on an hourly, as-needed basis.
- Can consider on-site, off-site, or offshore resources to address a spectrum of budget concerns.
- Flexible, scalable solution to address project demands.
- Learn More about DSaaS
11/13/19: Analytics at the Speed of Open Source