-
Quality Attributes for ML-Enabled Systems: A Research Synthesis
A synthesis of 24 academic papers examining quality attributes for ML-enabled systems -- from code quality and testing to fairness, explainability, and supply architectural tactics.
-
Engineering LLM-Based Agentic Systems
An exploration of LLM-based agentic systems -- their definitions and taxonomies, architectural patterns for reasoning and planning, responsible guardrails, and interoperability standards.
-
Insights into Recent MLOps practices
A systematic analysis of 23 articles on MLOps practices -- covering challenges in automation and tooling, model maintenance and artifact versioning, Responsible AI integration, and underexplored areas like scale, cost management, and foundation model operations.
-
Insights into Recent Engineering Practices for Machine Learning
An analysis of emerging ML engineering practices across five domains -- requirements engineering, architectural design, AutoML, data engineering, and Responsible AI -- examining new developments, persistent gaps, and their relation to established practice catalogues.
-
Machine Learning Engineering Practices in Recent Years: Trends and Challenges
A follow-up study analyzing 108 articles from 2022-2025 on ML engineering practices -- identifying five major research directions including MLOps, agentic architectures, and Responsible AI, alongside three underrepresented areas in data engineering, scaling, and user feedback.
-
Why do ML engineers struggle to build trustworthy ML applications?
Adoption of practices for trustworthy ML.
-
AutoML adoption in software engineering for machine learning
Adoption of AutoML and other practices for applications with ML components.
-
The 2020 State of Engineering Practices for Machine Learning
We published the results of our global survey on the adoption of engineering practices for ML teams.
-
A catalogue of Engineering best practices for Machine Learning
We published a catalogue of engineering best practices for ML applications.
-
A global survey on SE4ML
Measuring the adoption of engineering best practices for machine learning.
