This presentation anchors best practices for Enterprise Data Science based on Microsoft's "Team Data Science Process". The talk includes introducing the concepts, describing some real-world advice for project planning, and discusses typical titles of professionals who make enterprise data science successful. These techniques also apply for AI (artificial intelligence), deep learning, machine learning, and advanced analytics.

1. Managing Enterprise Data
Science
• Mark Tabladillo Ph.D.
• Cloud Solution Architect
• Microsoft
• April 12, 2019
This Photo by Unknown Author is licensed under CC BY-SA-NC

4. Domain specific pretrained models
To simplify solution development
Azure
Databricks
Machine
Learning VMs
Popular frameworks
To build advanced deep learning solutions
TensorFlowPyTorch ONNX
Azure Machine
Learning
LanguageSpeech
…
SearchVision
Productive services
To empower data science and development teams
Powerful infrastructure
To accelerate deep learning
Scikit-Learn
Familiar Data Science tools
To simplify model development
CPU GPU FPGA
From the Intelligent Cloud to the Intelligent Edge
Azure Notebooks JupyterVisual Studio Code Command line

5. Organization Reduces Risk

6. Organization Reduces Lifetime
Cost

7. Outline
Examine Each Phase of Team
Data Science Process
Examine Project Management
Examine Project Personas

8. TDSP Phases

9. Data Science Lifecycle
•Primary phases:
Lifecycle

10. Data Science Lifecycle
•Primary phases:
Lifecycle

11. Business Understanding
Study AI
Applications
Assess AI Maturity Identify Use Cases
and Personas

12. Data Acquisition and Understanding
Identify Possible
Data Sources
Identify Subject
Matter Experts
Evaluate ETL
Technologies

13. Modeling
Evaluate Modeling
Technologies
Feature
Engineering
Evaluate Modeling

14. Deployment
Standardize
Repositories
Persona
Delivery
CI/CD Strategy

15. TDSP Project Planning

16. Business Understanding
Data Preparation
Modeling
Deployment
Time

17. Business Understanding
Data Preparation
Modeling
Deployment
Time
Business Understanding

18. Time
Science One
Envisioning
Scoping
Science Two
Science Three
Science Four

19. TDSP Personas

20. Team Data
Science
Process –
Personas
Persona Needs
Account Executive • Effective planning and expectation setting with CxO level, directors, technical
leaders
• Integrating TDSP into a solutions conversation
Program
Manager/Engagement
Manager
• Resource Planning and Assignment
• Project Progress Tracking
• Communicating an effective narrative throughout the project lifecycle
Solution Architect • Architecture Development
• Technology Selection
• Central role of Data as a TDSP process driver
Data Scientist/Data
Engineer
• Understanding the major categories of the TDSP
• Understanding the latitude for choosing technologies
• Understanding the ability to generate and manage tasks under the major TDSP
categories
Application Developer • Integrating data science into a new or preexisting application
• Understanding the two types of developer roles
• Those directly participating in the TDSP iterations
• Those just providing input and consuming output
DevOps • Understanding how TDSP may be invoked in a production environment, either on
a planned or unscheduled basis
• How to work effectively with data scientists
• Optimizing and improving monitoring of production models

21. Resources
• Team Data Science Process https://docs.microsoft.com/en-
us/azure/machine-learning/team-data-science-process/
• Azure DevOps https://docs.microsoft.com/en-
us/azure/devops/?view=azure-devops
• AI Business School
https://www.microsoft.com/en-us/ai/ai-business-school

22. Summary
Examine Each Phase of Team
Data Science Process
Examine Project Management
Examine Project Personas

23. Connect with
Mark
Tabladillo
Linked In
Twitter @marktabnet