About me
What I Do
I build applied AI, data, and software systems for infrastructure teams, with a focus on drinking water, internal platforms, and practical decision support. My work sits between product discovery and implementation: understanding the operational context, shaping the technical architecture, and helping people adopt tools that are useful enough to survive beyond a pilot.
I am deliberately broad across frontend, backend, data, AI, and domain modeling, but the common thread is consistent: turning physical-infrastructure knowledge into software systems that can be inspected, trusted, and maintained.
At Dunea, I work on internal platforms, AI adoption, reusable Python tooling, digital-twin initiatives, web applications, dashboards, and engineering practices for data teams. A recent digital-twin platform effort pulled many of these threads together: operational telemetry, lab data, treatment context, source metadata, topology, scenarios, replay, reporting, validation logic, and paths toward solver-backed analysis.
I am especially interested in the gap between promising technology and organizational reality: maintainability, trust, ownership, governance, and the daily habits that decide whether software becomes infrastructure or just another experiment.
Background
My path started in civil engineering and water management at Delft University of Technology. Fieldwork in Thailand and Indonesia gave me a lasting respect for the physical, social, and institutional realities behind infrastructure problems. During my master's degree, I moved deeper into programming, data analysis, machine learning, and software development.
That combination still shapes how I work. I care about models and code, but also about what happens before and after them: data quality, domain language, workflows, incentives, documentation, and the ability for other people to understand and maintain what has been built.
How I Frame The AI Part
I use AI to amplify domain knowledge and software judgment, not to replace them. The useful part is not that AI can generate code quickly; it is that it can compress implementation cycles when the human steering the work understands the domain, the architecture, and the validation burden.
Read more about my operating styleAreas I Keep Coming Back To
- Applied AI that fits real governance, risk, and operational constraints
- Internal platforms that make good engineering easier for data teams
- Decision-support tools for water systems, planning, and operations
- Digital twins, solver integration, simulation environments, and scenario analysis
- AI-augmented work that improves judgment instead of hiding it
Research & Public Context
Some of my work has public reference points, even when the strongest implementation details remain internal. I contributed substantial domain, product, and implementation input to KWR/BTO roadmap work on digital twins for drinking-water treatment, and I presented a practical generative-AI use case in a KWR hydroinformatics knowledge-exchange setting.
- KWR / BTO digital twin roadmap for drinking-water treatment
- KWR hydroinformatics article on generative AI in water utilities
- MSc thesis: Investigating the MetOp ASCAT vegetation parameters
- TU Delft field project: mobile water monitoring in the Phetchaburi River Basin
- BSc thesis fieldwork: improving water treatment processes in Bandung
Let's talk
I'm interested in serious conversations about digital water, internal platforms, AI-assisted engineering, and decision-support systems for infrastructure teams.