MLOps applies agile and DevOps practices specifically to machine-learning projects, connecting research implementations with the constraints of production systems. section 1.3
Use MLOps to structure delivery around explicit project concerns:
- Set up the project by clarifying goals, user experience, performance constraints, evaluation, personalization, people, compute, and infrastructure. section 1.3
- Treat data pipelines as delivery concerns, including privacy and bias. section 1.3
- Contrast research needs with production requirements, including performance and compute considerations. section 1.3
Figure 1.4 ML project flow as described in Huyen and Hopper (2020). Figure redrawn and adapted by the author. Arrows represent dependency relations rather than workflow.
Explore the ML project flow and its dependency relationships across delivery phases. section 1.3
