Man, class of 2026 and M2 student in Data and Artificial Intelligence, is completing a research internship at DeVinci Research Center. His work focuses on DaMoOp, a framework that connects data modelling decisions with performance, cost, and environmental impact.
This experience reflects current developments in data engineering, where technical choices increasingly incorporate measurable, multidimensional criteria.
A research internship at the intersection of data and AI at ESLV
At DeVinci Research Center, Man Patel contributes to research activities related to data systems and artificial intelligence. His internship focuses on the analysis and implementation of DaMoOp, a methodology introduced in 2025 in the journal Information Systems.
This framework addresses a recurring question in large-scale data architectures: how to select the most relevant data model depending on usage. Traditional approaches often rely on experience or isolated benchmarks. DaMoOp proposes a structured alternative based on automated generation and comparison of models.
Within this context, the work carried out during the internship combines theoretical research and applied development. The objective is to design tools that support architectural decisions with quantifiable indicators.
From conceptual models to automated data model selection
DaMoOp starts from a conceptual representation of data and generates several possible logical models. These models can correspond to relational databases, document-oriented systems, or column-based storage solutions.
The originality of the approach lies in the evaluation process. Each generated model is assessed through a multidimensional cost framework, structured around three main criteria:
- Time cost, linked to query latency and system performance
- Financial cost, related to cloud infrastructure and resource consumption
- Environmental cost, measured through carbon footprint indicators
This comparative analysis allows engineers to move from intuition-based decisions to measurable trade-offs. It also introduces a structured way to anticipate the long-term implications of data architecture choices.
Integrating sustainability into data architecture
One of the key aspects of the research conducted during the internship concerns environmental impact. Data infrastructures rely on energy-intensive systems, and optimisation strategies often focus on performance or financial efficiency.
The work carried out within DaMoOp integrates environmental metrics directly into the decision-making process. This perspective leads to new questions: how does data structure influence energy consumption? How can query optimization reduce carbon emissions?
Man Patel contributes to the development of tools that quantify these aspects. By correlating query patterns, data evolution, and infrastructure usage, the research aims to identify configurations that balance efficiency and sustainability.
Research topics explored during the internship
Several research directions are addressed as part of this internship at DeVinci Research Center:
The first focuses on model obsolescence. The objective is to understand when NoSQL data models become less relevant as usage patterns evolve, and to define indicators that trigger schema migration.
Another axis concerns recommendation systems for data modelling. The work involves designing tools that automate the selection of low-carbon data models within frameworks such as DaMoOp and FACT-DM.
Performance benchmarking also plays a central role. By simultaneously analysing query latency, cloud billing, and environmental impact, the research provides a broader view of system optimisation.
These activities contribute to a more comprehensive understanding of data systems, where technical performance is evaluated alongside economic and environmental dimensions.
A learning experience aligned with current industry challenges
This internship reflects ongoing transformations in data engineering. The increasing volume of data and the complexity of distributed systems require new methodologies to guide architectural decisions.
The integration of sustainability metrics introduces additional constraints, but also new opportunities for innovation. Engineers are now expected to consider the broader impact of their technical choices.
As part of the ESILV engineering curriculum, this research experience offers exposure to both academic work and applied problem-solving. It connects theoretical models with real-world challenges encountered in large-scale data systems.
Research strategy: excellence, interdisciplinarity and global outlook
ESILV’s research strategy is structured around four main development objectives: scientific excellence, societal impact, interdisciplinarity, and internationalisation.
Scientific excellence is reflected in the growing number of publications in leading academic journals, combined with dissemination efforts through practice-oriented contributions and accessible formats such as videos, as well as participation in contractual research projects, including chairs and European programs.
Interdisciplinarity is supported by the De Vinci Research Center, where researchers in engineering and management sciences collaborate in shared teams, enabling work on topics related to societal challenges such as the ecological transition and e-health.
Internationalisation also shapes this strategy, with the development of partnerships with foreign institutions, increased mobility of researchers, and the organisation of scientific events with an international reach.
Learn more about the Data Engineering & AI major at ESILV