4
Department of Machine Learning Systems, Royal Institute of Engineering, London, United Kingdom
4
Faculty of Data Science and Operations, University of Northwood, Toronto, Canada
Abstract
Background: The transition of machine learning (ML) and artificial intelligence (AI) models from research to production has exposed significant operational challenges. While Continuous Integration/Continuous Deployment (CI/CD) is a mature practice in traditional software engineering, its application in the ML lifecycle (MLOps) presents unique complexities, including data versioning, model retraining, and continuous monitoring. There is a notable gap in the literature regarding comprehensive case studies of successful, end-to-end CI/CD implementations for ML [31,32].
Objectives: This paper aims to identify the architectural patterns, key success factors, and best practices of successful CI/CD pipeline implementations for ML and AI systems through a comparative analysis of real-world case studies.
Methods: A qualitative, multiple case study methodology was employed. Data was systematically collected from publicly available, detailed accounts of CI/CD implementations from diverse industries, including e-commerce, healthcare, and finance. A thematic analysis framework was used to extract and compare key aspects such as pipeline architecture, toolchains, automation strategies, and measured outcomes.
Results: The analysis of the case studies revealed several common success patterns, including the extensive use of containerization, the adoption of centralized feature stores for managing ML-specific data, and the implementation of robust automated testing and validation stages beyond traditional code checks. Key differences were observed in pipeline design based on industry-specific requirements, such as regulatory compliance in healthcare and real-time inference demands in finance. Each case demonstrated a significant, measurable improvement in deployment velocity, operational stability, and model performance [34].
Conclusion: A well-architected CI/CD pipeline is a critical enabler for scaling ML and AI initiatives effectively. The findings from these case studies provide a practical framework and actionable insights for organizations seeking to build and refine their MLOps capabilities, moving from ad-hoc model deployment to a systematic, automated, and reliable process.
How to Cite
Michael Davis, & Prof. Evelyn Reed. (2025). Operationalizing MLOps: A Comparative Case Study of CI/CD Pipeline Implementations for AI and Machine Learning. Frontiers in Emerging Artificial Intelligence and Machine Learning, 2(10), 17–31. Retrieved from https://irjernet.com/index.php/feaiml/article/view/226
📄Bernardo, João Helis, et al. (2024). How do machine learning projects use continuous integration practices? An empirical study on GitHub Actions. 2024 IEEE/ACM 21st International Conference on Mining Software Repositories (MSR). IEEE.
📄Singh, Prerna. (2023). Systematic review of data- centric approaches in artificial intelligence and machine learning. Data Science and Management, 6(3), 144–157.
📄Johnston, Craig, & Johnston, Craig. (2020). In- platform CI/CD. In Advanced Platform Development with Kubernetes: Enabling Data Management, the Internet of Things, Blockchain, and Machine Learning (pp. 117–152).
📄Ratilainen, Katja-Mari. (2023). Adopting machine learning pipeline in existing environment.
📄Houerbi, Alaa, et al. (2024). Empirical analysis on CI/CD pipeline evolution in machine learning projects. arXiv preprint arXiv:2403.12199.
📄Mahida, Ankur. (2024). A review on continuous integration and continuous deployment (CI/CD) for machine learning.
📄Vemuri, Naveen, Thaneeru, Naresh, & Tatikonda, Venkata Manoj. (2024). AI-optimized DevOps for streamlined cloud CI/CD. International Journal of Innovative Science and Research Technology, 9(7), 10–5281.
📄Bagai, Rahul, Masrani, Ankit, Ranjan, Piyush, & Najana, Madhavi. (2024). Implementing continuous integration and deployment (CI/CD) for machine learning models on AWS.
📄Liang, Penghao, et al. (2024). Automating the training and deployment of models in MLOps by integrating
📄systems with machine learning. arXiv preprint
📄arXiv:2405.09819.
📄Bernardo, João Helis, et al. (2024). How do machine learning projects use continuous integration practices? An empirical study on GitHub Actions. 2024 IEEE/ACM 21st International Conference on Mining Software Repositories (MSR). IEEE.
📄Makani, Sai Teja, & Jangampeta, ShivaDutt. (2024). The evolution of CI/CD tools in DevOps from Jenkins to GitHub Actions.
📄Paguthaniya, Sajid Ali, et al. (2024). Integration of machine learning models into backend systems: Challenges and opportunities.
📄Vemulapalli, Gopichand. (2023). Operationalizing machine learning best practices for scalable production deployments. International Machine Learning Journal and Computer Engineering, 6(6), 1– 21.
📄Pandi, Srinivas Babu. (2023). Artificial intelligence in software and service lifecycle.
📄Brandon, Colm, & Margaria, Tiziana. (2023). Low- code/no-code artificial intelligence platforms for the health informatics domain. Electronic Communications of the EASST, 82.
📄Siltala, Ville. (2023). Machine learning operations architecture in healthcare big data environment: Batch versus online inference (Master’s thesis).
📄Lähteenmäki, Jaakko, et al. (2023). Agile and holistic medical software development: Final report of AHMED project.
📄Makarov, Vladimir, et al. (2024). Good machine learning practices: Learnings from the modern pharmaceutical discovery enterprise. Computers in Biology and Medicine, 177, 108632.
📄Theusch, Felix, et al. (2023). Towards machine learning-based digital twins in cyber-physical systems. AI4DT&CP@ IJCAI.
📄Shankar, Shreya, et al. (2024). "We have no idea how models will behave in production until production": How engineers operationalize machine learning. Proceedings of the ACM on Human-Computer Interaction, 8(CSCW1), 1–34.
📄Deutsch, Daniel. (2023). Machine learning operations–domain analysis, reference architecture, and example implementation. LL.B.(WU), LL.M.(WU).
📄Chandra, R., Lulla, K., & Sirigiri, K. (2025). Automation frameworks for end-to-end testing of large language models (LLMs). Journal of Information Systems Engineering and Management, 10(43s), e464–e472. https://doi.org/10.55278/jisem.2025.10.43s.8400
📄Chandra, R., Bansal, R., & Lulla, K. (2025). Benchmarking techniques for real-time evaluation of LLMs in production systems. International Journal of Engineering, Science and Information Technology, 5(3), 363–372.
📄Sirigiri, Karthik, Chandra, Reena, & Lulla, Karan. (2025). Impact of cloud-native CI/CD pipelines on deployment efficiency in enterprise software. International Journal of Computational and Experimental Science and Engineering, 11(2). https://doi.org/10.22399/ijcesen.2383
📄Durgam, S. (2025). CICD automation for financial data validation and deployment pipelines. Journal of Information Systems Engineering and Management, 10(45s), 645–664.
📄Lulla, K. (2025). Python-based GPU testing pipelines: Enabling zero-failure production lines. Journal of Information Systems Engineering and Management, 10(47s), 978–994.
📄Venkiteela, P. (2025). Modernizing opportunity-to- order workflows through SAP BTP integration architecture. International Journal of Applied Mathematics, 38(3s), 208–228. https://doi.org/10.58298/ijam.2025.38.3s.12
📄Gannavarapu, P. (2025). Performance optimization of hybrid Azure AD join across multi-forest deployments. Journal of Information Systems Engineering and Management, 10(45s), e575–e593. https://doi.org/10.55278/jisem.2025.10.45s.575
📄Koneru, N. M. K. (2025). Centralized logging and observability in AWS: Implementing ELK stack for enterprise applications. International Journal of Computational and Experimental Science and
📄Koneru, N. M. K. (2025). Leveraging AWS CloudWatch, Nagios, and Splunk for real-time cloud observability. International Journal of Computational and Experimental Science and Engineering, 11(3). https://doi.org/10.22399/ijcesen.3781
📄Hariharan, R. (2025). Zero trust security in multi- tenant cloud environments. Journal of Information Systems Engineering and Management, 10(45s). https://doi.org/10.52783/jisem.v10i45s.8899
📄Reddy Dhanagari, M. (2025). Aerospike: The key to high-performance real-time data processing. Journal of Information Systems Engineering and Management, 10(45s), 513–531.
📄Chandra, R. (2025). Reducing latency and enhancing accuracy in LLM inference through firmware-level optimization. International Journal of Signal Processing, Embedded Systems and VLSI Design, 5(2), 26–36. https://doi.org/10.55640/ijvsli-05-02-02
📄Bonthu, C., Kumar, A., & Goel, G. (2025). Impact of AI and machine learning on master data management.
📄Malik, G., Rahul Brahmbhatt, & Prashasti. (2025). AI- Driven Security and Inventory Optimization: Automating Vulnerability Management and Demand Forecasting in CI/CD-Powered Retail Systems. International Journal of Computational and Experimental Science and Engineering, 11(3). https://doi.org/10.22399/ijcesen.3855
📄Prassanna R Rajgopal. (2025). AI-optimized SOC playbook for Ransomware Investigation. International Journal of Data Science and Machine Learning, 5(02), 41-55. https://doi.org/10.55640/ijdsml-05-02-04
📄Evaluating Effectiveness of Delta Lake Over Parquet in Python Pipeline. (2025). International Journal of Data Science and Machine Learning, 5(02), 126-144. https://doi.org/10.55640/ijdsml-05-02-12
