A quick article to explain the difference between two relatively new concepts: MLOps and DataOps.
DataOps
DataOps is an automated, process-oriented methodology, used by analytic and data teams, to improve the quality and reduce the cycle time of data analytics. (Wikipedia)
What is inside?
- Methods from DevOps
- Deployment of data pipelines such as data extraction and data transformation
- Monitoring for data pipelines
Target
Data teams such as data analytics and data engineering teams.
Keywords
- Data
- Pipeline
- DevOps
- Deployment
- Monitoring
MLOps
MLOps (a compound of “machine learning” and “operations”) is a practice for collaboration and communication between data scientists and operations professionals to help manage production ML […] lifecycle. (Wikipedia)
What is inside?
- Methods from DevOps
- Deployment of machine learning models
- Monitoring of machine learning models
- Reproducibility of machine learning models
- Scalability of machine learning models
Target
- Data scientists with data engineers, operational workers and business teams
Keywords
- Machine learning
- Data science
- Deployment
- Monitoring
- DevOps
Conclusion
In summary, DataOps is for deploying and monitoring data pipelines.
MLOps is for deploying and monitoring machine learning models.
You can have DataOps without MLOps because you can have data extraction and transformation without machine learning. The contrary is rarely true.
Thank you for reading. Feel free to contact me on Twitter if you want to discuss that.