Interviewed by: Mehadi Hasan
Developer: Fady Desoky Saeed Abdelaziz
Role: Data Analyst | Data Analytics & Data Engineering | Enterprise Systems & Process Optimization
Location: Cairo, Egypt
Software engineering, data analytics, and research are often treated as separate disciplines. For Fady Desoky Saeed Abdelaziz, however, they are closely connected by a common question: how can technical information be transformed into something people can understand and use to make better decisions?
Fady's experience spans backend development, software testing, business process optimization, enterprise reporting, machine learning, and applied research. These experiences have shaped an approach to engineering that goes beyond implementation. He is interested in understanding the problem behind a system, evaluating the reliability of its data, and presenting technical findings in a way that supports practical decisions.
That perspective is evident in his projects. Qualioro explores how software quality metrics can become understandable recommendations. His Retail FMCG Sales Analytics project connects transaction data, predictive modelling, dashboards, and an interactive application. His distributed-systems work investigates how services behave when failures occur, while his research examines the relationship between computing resources and energy consumption.
Across these projects, a consistent theme emerges: technical measurements become more valuable when their meaning, limitations, and implications are made clear.
A Journey Across Software Engineering and Data
Fady's journey began with his studies in Software Engineering at Sadat Academy for Management Sciences. He was interested in understanding how software systems work, how applications are designed, and how technology can solve practical problems.
His early experience with backend development and software testing introduced him to application logic, database interactions, and reliability. These foundations helped him understand the importance of building systems that behave correctly and remain maintainable.
His subsequent work in business process optimization and data analytics expanded that perspective. Instead of considering software only in terms of its implementation, he began examining how systems operate within larger organizations, how workflows affect the people who use them, and how data influences operational decisions.
Research added another dimension. His work on sustainable software engineering encouraged him to investigate computing resource consumption, while his urban mobility research introduced him to the challenges of analysing real-world location data.
Each area contributed a different way of approaching technical problems. Software engineering emphasizes architecture, correctness, and maintainability. Process optimization focuses on workflows and inefficiencies. Data analysis encourages evidence-based reasoning and the validation of assumptions.
As Fady explains, working across these disciplines has helped him avoid jumping straight into implementation before understanding the actual problem.
This approach also influences how he develops new skills. Rather than studying technologies in isolation, he prefers applying them to practical projects that connect with his existing experience and the problems he wants to solve.
What Enterprise Operations Taught Him About Data Reliability
One of the most influential parts of Fady's experience has been his work with Nestlé Business Solutions across HR Operations Process Optimization and Data Analytics & Reporting.
In his earlier process optimization work, he contributed to a regional HR system migration initiative involving multiple countries. The work required an understanding of existing workflows, dependencies between systems, and the consistency of historical employee records.
His profile also describes involvement in the MENA Gratuity Project, a regional migration initiative across six Gulf Cooperation Council countries. His responsibilities included mapping workflow dependencies, identifying data gaps and potential migration risks, documenting processes, and supporting historical data validation.
These activities illustrate a practical challenge in enterprise technology: moving information between systems is not simply a matter of transferring records. The information must remain accurate and consistent as processes and systems change.
Later, his work in data analytics and reporting involved business data used in dashboards and operational reporting across regional markets. His profile describes reporting frameworks and data workflows supporting operations across more than ten MENA and Gulf markets, alongside reporting activities involving more than 60 retail partners.
Together, these experiences reinforced an important lesson about reliability.
A system can function correctly from a technical perspective and still produce unreliable results if its underlying data is inaccurate, incomplete, or inconsistent. An error in a source record or transformation can propagate into a report and potentially influence a business decision.
For Fady, data validation, process documentation, and consistency checks are therefore essential parts of reliable systems, not secondary tasks to be considered after implementation.
This principle carries through to his independent projects, where he places particular emphasis on explaining what data represents and what conclusions can reasonably be drawn from it.
Qualioro: Making Software Quality Metrics More Useful
Software quality tools can generate measurements that indicate potential weaknesses in a codebase. However, a score by itself does not necessarily explain what caused the result, why it matters, or what a developer should investigate first.
That gap motivated Fady to build Qualioro, a software quality intelligence prototype designed to connect measurable indicators with understandable explanations and targeted recommendations.
Qualioro evaluates indicators such as bug density, code complexity, test coverage, and team output. It then presents an overall quality score, a risk classification, a confidence indicator, and recommendations based on weaker areas.
The idea is to move beyond simply displaying metrics. Developers should be able to explore the factors contributing to an assessment and identify areas that may deserve further attention.
Fady built the application using Next.js 15, TypeScript, Tailwind CSS, shadcn/ui, and Recharts. Its workflow separates the evaluation of input metrics from the interpretation of results and their presentation in an interactive interface.
The interface includes quality scoring, risk breakdowns, visualizations, and a what-if simulation. By adjusting inputs, users can observe how the overall assessment changes. An advanced view also exposes more of the scoring breakdown, making the result easier to inspect.
These features reflect a central design consideration: an assessment is more useful when users can understand how it was produced rather than being asked to accept a number without context.
Explainability Without False Confidence
One of the challenges in designing Qualioro is balancing simplicity with accuracy.
A simplified scoring model can demonstrate how different quality indicators might contribute to an assessment. However, its output should not be mistaken for a definitive evaluation of an actual software repository.
Fady makes this distinction clear in the current implementation. The analysis runs entirely on the client side, without sending or storing user data through an external service. The model uses simplified inputs to demonstrate the concept; it does not automatically analyse arbitrary repositories as a production-grade system.
He sees real repository integrations, automated metric extraction, CI/CD data, and broader evaluation as possible future extensions.
His approach to explainability can be summarized through three questions:
- What was detected?
- Why might it matter?
- What should the developer investigate next?
These questions provide a useful framework for turning technical measurements into actionable information.
The broader lesson is that transparency is not simply a matter of presenting information clearly. It also means communicating uncertainty and avoiding claims that the available evidence cannot support.
Understanding Failure in Distributed Systems
Fady's interest in reliability also extends to distributed applications.
His Mobile Cloud System project explores a simulated cloud-native microservices architecture through an interactive observability dashboard. Rather than treating a distributed application as a single unit, the project examines how multiple services communicate, how failures can affect connected components, and how developers can understand what happens when a workflow breaks down.
The project explores several important distributed-systems concepts, including the Saga pattern, retries, circuit breakers, and observability.
Each addresses a different aspect of reliability. Retries can help recover from temporary errors. Circuit breakers can prevent repeated calls to a service that is experiencing problems. Distributed transaction patterns such as Saga help coordinate operations across service boundaries.
These mechanisms are important because distributed applications introduce failure points that are less straightforward than those in a single component. A service may become unavailable, communication may fail, or an error in one part of a workflow may affect subsequent operations.
For Fady, the project reinforced a fundamental engineering principle: reliable distributed systems must be designed with failure in mind rather than assuming every request will succeed.
Observability is equally important. When a workflow fails, developers need enough information to identify where the failure occurred, understand its effects on other services, and determine whether the system recovered as expected.
The project provides a way to explore these behaviours in a simulated environment. Its value lies in making distributed-system interactions and failure-handling concepts easier to examine.
Together with Qualioro, it demonstrates Fady's interest in building tools that help developers understand not just what a system does, but also how and why it behaves in particular ways.
From 100,000 Retail Transactions to Business Insights
Fady's Retail FMCG Sales Analytics project approaches a different problem: how to transform a large transaction dataset into information that can support business decisions.
The project uses a dataset containing 100,000 transactions and 21 variables covering products, customers, revenue, costs, stores, sales channels, payment methods, and inventory.
Rather than focusing on a single analytical technique, Fady combined data preparation, exploratory analysis, visualization, predictive modelling, and an interactive application into one workflow.
He used Python, Pandas, and NumPy for data preparation and analysis, Scikit-learn for machine learning, Power BI for interactive dashboards, and Streamlit to make the analytical workflow accessible through an application.
The analysis explored questions such as which product categories and brands performed best, how sales channels compared, what patterns appeared in customer behaviour, and where inventory or profitability issues might warrant further investigation.
The project also included a Linear Regression model designed to predict margin using available business and transactional features.
The documented evaluation results were:
- R²: 0.863
- Mean Absolute Error (MAE): 19.28
- Root Mean Squared Error (RMSE): 27.64
These metrics describe different aspects of the model's performance. R² indicates how much of the variation in the target variable is explained by the model under the evaluation setup, while MAE and RMSE measure prediction errors using different approaches.
The figures provide an evaluation of this particular model and dataset; they should not be interpreted as a guarantee of performance on other retail datasets or in other business environments.
For Fady, one of the project's most valuable aspects was connecting the stages of the analytical process. Data preparation established a foundation for analysis, exploratory work helped reveal patterns, dashboards made the results easier to investigate, and predictive modelling demonstrated how a model could become part of an interactive workflow.
The underlying lesson is that analytics is most useful when technical methods remain connected to a clear business question and the results are presented in a form people can understand and use.
The project is available on GitHub.
Another important area of Fady's work is sustainable software engineering.
Software depends on physical computing infrastructure, and its execution consumes resources through components such as CPUs, GPUs, and memory systems. Understanding these demands can help engineers investigate resource efficiency alongside more familiar concerns such as correctness and performance.
Fady's research, Analyzing Software Resource Consumption Data for Sustainable Green Software Development: A Predictive Model to Reduce and Optimize Energy Consumption in Software, examines the relationship between resource-consumption variables and energy usage using a dataset covering 2020–2025.
The project reflects his interest in whether resource-consumption data can help explain and predict energy use, potentially supporting more informed engineering decisions.
This work connects software engineering with environmental responsibility. Developers may not control every aspect of the infrastructure on which their applications run, but they can measure resource consumption, investigate inefficiencies, and evaluate the effects of potential optimizations.
Fady's profile identifies him as the first author and presenter of this research at the AII 2025 International Conference in Washington, DC. The abstract was published in the conference proceedings, with the full paper described as being under preparation for publication through Springer Nature.
His interest in sustainability also complements his work on software quality and distributed systems. Each involves understanding system behaviour through evidence rather than relying solely on assumptions or surface-level indicators.
In software quality, the objective is to interpret metrics responsibly. In distributed systems, it is to understand failures and recovery. In sustainable software engineering, it is to investigate the relationship between resource use and energy consumption.
Across all three, measurement is a starting point for investigation rather than an end in itself.
Learning from Real-World Data: Urban Mobility Research
Fady's research interests also extend beyond software systems into urban mobility.
During his research internship at Nile University, he worked on GNSS-based urban traffic modelling using real-world GPS traces from urban traffic networks. The work involved processing large-scale mobility datasets and applying statistical and time-series modelling to forecast traffic congestion patterns.
He co-authored Graphing the Jam: GNSS-Powered Modeling and Forecasting of Urban Traffic Congestion, which was presented at the NetMob 2025 Data Challenge & Conference in Paris, France.
Working with real-world location data introduces challenges that differ from those encountered with simplified examples. The data must be processed and analysed carefully before it can support useful conclusions about traffic patterns.
This research complements Fady's wider interest in transforming complex datasets into predictive insights. Whether the subject is software resource consumption, retail transactions, or urban traffic, the analytical challenge is similar: identify meaningful patterns, evaluate them using appropriate methods, and communicate the results without overstating what the evidence establishes.
The publication is available through ResearchGate.
A Consistent Engineering Philosophy
Although Fady's projects cover different domains, several principles connect them.
Understand the problem before choosing the implementation. His experience in process optimization taught him to examine existing workflows, dependencies, and the people affected by a system before jumping into development.
Treat data quality as part of system quality. Enterprise reporting and migration work demonstrated how inaccurate or inconsistent information can undermine otherwise functional systems.
Make technical results understandable. Qualioro explores how measurements can be connected to explanations and recommendations, while the Retail FMCG project brings analytical results into dashboards and an interactive application.
Design for failure and uncertainty. His Mobile Cloud System project explores failure-handling mechanisms and observability, while Qualioro emphasizes the limitations of simplified scoring models.
Connect technical work to practical outcomes. Across analytics and research, Fady aims to investigate questions that matter beyond the implementation itself, whether they concern business performance, system reliability, or energy consumption.
These principles reflect an approach to engineering in which the technology is only one part of the solution. Understanding the context, validating the evidence, and communicating the results are equally important.
What Comes Next
Fady's current interests include sustainable and green software systems, environmental data analysis, predictive modelling, data engineering, and cloud-based data systems.
His projects also suggest several directions for further exploration. Qualioro could be extended through real repository integrations and automated metric extraction. The Retail FMCG analytics workflow demonstrates how data preparation, machine learning, and interactive reporting can work together. His sustainability research investigates how computing resource measurements may contribute to better-informed decisions about software efficiency.
These areas provide opportunities to deepen the connection between software engineering, data, and measurable outcomes.
Rather than treating software development, analytics, and research as isolated activities, Fady's work brings them together around a shared objective: understanding complex systems and making the information they produce more useful.
The result is an engineering perspective that values not only building a system, but also examining its behaviour, validating its outputs, and helping people make informed decisions based on what the evidence actually shows.
About Fady Desoky Saeed Abdelaziz
Fady Desoky Saeed Abdelaziz is a Cairo-based data analyst with a background in software engineering. His interests span data analytics, data engineering, enterprise systems, process optimization, machine learning, sustainable software engineering, and applied research.
His projects include Qualioro, a software quality intelligence prototype; Mobile Cloud System, a simulated microservices architecture with an observability dashboard; and Retail FMCG Sales Analytics, which combines transaction analysis, predictive modelling, Power BI, and Streamlit. His research includes sustainable software resource consumption and GNSS-based urban traffic modelling.
Developer Stories is a CoderLegion interview series highlighting the people, ideas, and engineering decisions behind software development.
