Data Science in the Water Sector · EngD
Akrivi Alexandraki
I build tools that help water utilities make decisions with their data, and I study how to keep machine-learning models reliable in daily use.
For my TU Delft Engineering Doctorate, I built and deployed a drinking-water quality platform for Waternet’s Leiduin plant.
Selected projects
I’ve worked with water-quality measurements, river discharge, weather observations and lab data. These projects show how I use each to answer an engineering question.
Research and publications
My doctoral work connects platform development with questions about model reliability. Earlier research covered European river flows and microplastics in sewage sludge.
Adapting FMECA for digital reliability assessment of soft sensors in drinking-water systems
Alexandraki, Kyritsakas, van den Boomen and Rietveld · Presented at HIC 2026
Development of a Digital Twin-Based Decision Support Platform for Drinking Water Quality Monitoring and Management
Alexandraki, Kyritsakas, van der Helm, Kossieris, Makropoulos and Rietveld
Trend analysis of the runoff patterns in Central and Northern Europe: Identification of changes in magnitude and timing
Changes in the magnitude and timing of river flows across five countries.
Microplastics in sewage sludge destined to anaerobic digestion: The potential role of thermal pretreatment
Cesaro, Pirozzi, Zafırakou and Alexandraki
About me
I’m a civil engineer working on digital tools for water infrastructure.
I completed my Engineering Doctorate at TU Delft cum laude. In the Horizon Europe ToDrinQ project, I worked with Waternet and research partners to bring live measurements and machine-learning models into a platform for drinking-water quality.
My background includes an MSc in Civil Engineering focused on water management and hydrology at TU Delft and a Civil Engineering diploma from Aristotle University of Thessaloniki. I’ve also led development of a tool for comparing industrial decarbonisation scenarios at the Port of Rotterdam.
Tools and methods
Data analytics
My environmental data work provides a foundation for exploring people analytics: understanding patterns, questioning data quality and communicating evidence.
Statistical thinking
Explore runoff trend analysis and the questions behind interpreting change over time.
Data to decision support
See how the Leiduin platform connects live measurements, model outputs and operational decisions.
Reliability & oversight
See how I assess failure modes and consider detection, mitigation and oversight.
Where does the hiring funnel lose momentum?
This portfolio demo uses synthetic hiring data to illustrate funnel reporting.
Metric definitions & interpretation
Funnel: a completed application cohort, counted once per application at each stage reached. Bar lengths show the share of applications reaching each stage. Stage conversion uses the preceding stage as its denominator.
Offer acceptance: accepted offers divided by offers made. Application-to-hire: accepted offers divided by applications, assuming all accepted candidates join in this example.
Compare equivalent cohorts and role types before judging channel performance. A production dashboard would validate identifiers, missing stage dates and duplicate records, document refresh timing, and restrict access to confidential candidate information.
Work with me
I’m available for freelance work in digital water, applied AI and environmental data. Tell me a little about your project.