Home Research Hira Naveed
PhD Student · Associate Supervisor

Hira Naveed

Model-Driven Engineering for Runtime Monitoring of Human-Centric Requirements in ML Systems

HN

Monash University

Associate Supervisor: Dr Chetan Arora

Co-supervised with Prof. John Grundy · Dr Hourieh Khalajzadeh · Dr Omar Haggag

Research Theme: Responsible & Safe AI Systems →

Project illustration

Research Project

Model-Driven Engineering for Runtime Monitoring of Human-Centric Requirements in ML Systems

Hira's research works at the intersection of Model-Driven Engineering (MDE) and responsible AI, focusing on software systems that incorporate machine learning components. Her research addresses a critical gap: as ML systems are increasingly deployed in socially consequential settings, there is a growing need to monitor these systems at runtime for violations of human-centric requirements.

Human-centric requirements — such as fairness, privacy, and transparency — are qualitative and value-laden, making them difficult to specify formally and even more difficult to monitor automatically. Hira's work investigates how MDE techniques can provide the formal modelling and automated monitoring infrastructure needed to operationalise these requirements in live systems.

More recently, the research has extended to ML systems incorporating RAG pipelines, where the monitoring challenge is compounded by the dynamic, context-dependent nature of retrieval-augmented generation. The goal is a practical, model-driven monitoring framework that practitioners can adopt without deep expertise in formal methods — lowering the barrier to building trustworthy ML-enabled software.

Publications

Selected Publications.

IST · Journal 2024 Model driven engineering for machine learning components: A systematic literature review
+

Naveed, Hira, Chetan Arora, Hourieh Khalajzadeh, John Grundy, and Omar Haggag. Information and Software Technology 169 (2024): 107423. A thorough systematic literature review of MDE approaches applied to ML-enabled software systems — mapping the state of the field, identifying research gaps, and charting a path for future work.

MODELS · A 2024 Towards runtime monitoring for responsible machine learning using model-driven engineering
+

Naveed, Hira, John Grundy, Chetan Arora, Hourieh Khalajzadeh, and Omar Haggag. MODELS 2024. Presents an early framework for using model-driven engineering techniques to enable runtime monitoring of responsible AI properties in ML systems, with a focus on practical applicability.

ICSME · A 2025 Understanding Practitioners' Perspectives on Monitoring Machine Learning Systems
+

Naveed, Hira, John Grundy, Chetan Arora, Hourieh Khalajzadeh, and Omar Haggag. ICSME 2025. An empirical study investigating how practitioners currently approach monitoring of ML systems in production, surfacing challenges, tools, and unmet needs that inform the design of the proposed framework.

Collaboration

Interested in collaborating?

Get in Touch →