Dimitra Melanitis

Designed and delivered a full modernisation of Belgium’s meteorological encoding/decoding

infrastructure, transforming it into a reproducible, standards-compliant, test-driven data pipeline.

Functional Analysis & Requirements (October 2025):

● Conducted comprehensive functional and technical as-is analysis of meteorological data systems.

● Identified gaps, migration risks, and defined acceptance criteria for WIS 2.0 compliance.

● Analysed compatibility with international templates and WMO WIS 2.0 standards.

● Translated WMO international standards into functional specifications and test requirements.

● Defined upgrade pathways for AWS 10-minute and hourly SYNOP data streams.

● Defined migration roadmap from legacy libraries to Python 3.11 and eccodes.

● Drafted the implementation roadmap including code refactoring strategy, documentation approach,

and stakeholder coordination model.

● Delivered phased implementation plan with clear validation gates, quality checkpoints, dependencies,

and resource estimates to support migration programme.

DevOps / Engineering (November 2025):

● Built a containerised validation environment (Docker) enabling reproducible encoding/decoding tests.

● Integrated GitLab CI/CD pipelines for automated regression testing, BUFR compliance checks, and

performance profiling.

● Developed modular Python encoders/decoders with template-agnostic architecture supporting 8+

SYNOP variants.

● Implemented automated operational routing with safe-fail processing for invalid or incomplete files.

● Designed scalable architecture diagrams, data flow models, and API specs to support maintainability

and onboarding.

● Created an automated test harness comparing 70+ parameters with ECMWF reference outputs

(bufr_dump) and added tolerance-based numeric validation for unit conversions.

Designed a subscription management platform integrating multiple SaaS services and custom WordPress development to support the association's recurring revenue model.

I led the design, modernisation, and operation of large-scale, mission-critical data systems in a highly regulated environment.

As DevOps Lead and Solution Architect, I replaced a decades-old Fortran-based national data dissemination system with a modern, automated, and fully traceable platform processing several terabytes of data per day. I carried out end-to-end system analysis (as-is / to-be), defined migration roadmaps, and coordinated IT, scientific, and administrative stakeholders. I designed CI/CD pipelines enabling automated testing, deployment, and safe production rollouts, and implemented event-driven automation (cron, inotify), observability (Zabbix, logging, health checks), and regression testing to ensure reliability, auditability, and operational resilience.

Over more than a decade, I designed and maintained automated data pipelines for product generation, conversion, and distribution (SQL, BUFR, GRIB, NetCDF), ensuring continuous compliance with international standards (WMO, ECMWF). This included time-critical processing, operational validation, and continuous optimisation for performance and stability.

In parallel, I acted as Business/Technical Analyst and Product Owner for workflow digitisation projects, replacing paper and legacy processes with automated, traceable Jira Service Desk workflows, including document generation and lifecycle validation.

I also served as Belgium’s international data-exchange coordinator, managing secure data transfers, API and HPC/cloud access, and ensuring interoperability and continuity with European partners. I delivered training to external organisations, including the Belgian Armed Forces.

In 2024, I led the migration of an operational forecasting model from on-premise HPC infrastructure to ECMWF’s ATOS supercomputer in Bologna, coordinating cross-organisation teams, validation phases, and progressive rollouts to guarantee service continuity and reproducibility.