If the broader topic of product development "blends the perspective of marketing, design, and manufacturing into a single approach to product development,"[6] then design is the act of taking the marketing information and creating the design of the product to be manufactured.
The physical design relates to the actual input and output processes of the system. This is explained in terms of how data is input into a system, how it is verified/authenticated, how it is processed, and how it is displayed.In physical design, the following requirements about the system are decided.
Designing the overall structure of a system focuses on creating a scalable, reliable, and efficient system. For example, services like Google, Twitter, Facebook, Amazon, and Netflix exemplify large-scale distributed systems. Here are key considerations:
Machine learning systems design focuses on building scalable, reliable, and efficient systems that integratemachine learning (ML) models to solve real-world problems. ML systems require careful consideration of data pipelines, model training, and deployment infrastructure. ML systems are often used in applications such asrecommendation engines,fraud detection, andnatural language processing.
Key components to consider when designing ML systems include:
Problem Definition: Clearly define the problem, data requirements, and evaluation metrics. Success criteria often involve accuracy, latency, and scalability.[11]
Data Pipeline: Build automated pipelines to collect, clean, transform, and validate data.[12]
Deployment and Serving: Deploy trained models to production environments using scalable architectures such as containerized services (e.g.,Docker andKubernetes).[13]
Monitoring and Maintenance: Continuously monitor model performance, retrain as necessary, and ensuredata drift is addressed.[14]
Designing an ML system involves balancing trade-offs between accuracy, latency, cost, and maintainability, while ensuring system scalability and reliability. The discipline overlaps withMLOps, a set of practices that unifies machine learning development and operations to ensure smooth deployment and lifecycle management of ML systems.
^Werner, Ulrich (September 1987). "Critical heuristics of social systems design".European Journal of Operational Research.31 (3): 276-283.doi:10.1016/0377-2217(87)90036-1.
^Ulrich, Karl T.;Eppinger, Steven D. (2000).Product Design and Development (Second ed.). Boston: Irwin McGraw-Hill.
^Sorvisto, Dayne (2023).MLOps Lifecycle Toolkit: A Software Engineering Roadmap for Designing, Deploying, and Scaling Stochastic Systems. Apress.ISBN978-1-4842-9641-7.
^Polyzotis, Neoklis (2017). "Data Management Challenges in Production Machine Learning".Proceedings of the 2017 ACM International Conference on Management of Data. pp. 1723–1726.doi:10.1145/3035918.3054782.ISBN978-1-4503-4197-4.
This article incorporatespublic domain material fromFederal Standard 1037C.General Services Administration. Archived fromthe original on 2022-01-22.
This article incorporates text from this source, which is available under theCC BY 4.0 license.
