Research-grade web portal applying Divisible Load Theory to optimally schedule large, splittable workloads (big-data batches, parallel scans, rendering, simulations) across distributed processors, minimising both completion time and energy. C++/MILP solver backend with Gantt visualisation, time–energy Pareto analysis and HTTP/CLI access.
Python ETL and analytics pipeline (QUality AnalyzeR of application Efficiency) that ingests, stores and visualises performance-profiling data from HPC applications — currently focused on OpenFOAM CFD workloads with AMD uProf counters and MPI profilers.
Digital repository of 2+ million specimen records drawn from 200 years of natural history collections at the Faculty of Biology, AMU — covering algae, plants, fungi, lichens and animals. Built to ABCD and Darwin Core standards with online analytical tools, a mobile field application and GBIF data exchange.
Online portal that scores natural history collections against 20 transparent criteria and centralises management of documentation, loans, publications, citations and Nagoya-ABS permits. Built jointly with the Faculty of Biology, AMU; integrates with GBIF, BOLD and Crossref.
Web application that publishes datasets from internal CKAN portals directly to Zenodo, the CERN open-access repository — removing the manual download-and-re-upload step that institutions face when making research data citeable. Single-resource or full-dataset export, Keycloak SSO, asynchronous RabbitMQ transfers with retry, and live status tracking.
Python client that transfers large scientific datasets (multi-GB to TB) between CKAN data portals and HPC storage, letting files live on dedicated HPC infrastructure while metadata stays in CKAN — avoiding slow uploads through the CKAN web UI. SFTP and CKAN-API transports, SHA-256 integrity checks, automatic compression, hash-based deduplication, exponential back-off retries.
Go to top of the page.