Software, data and AI engineer in Brussels.
I train deep-learning models and build the apps and data pipelines around them.
jwds.me · LinkedIn · jawadshaissah@gmail.com
- MSc in Applied Sciences and Engineering: Applied Computer Science, Vrije Universiteit Brussel, 2026
- Thesis: a Residual U-Net that turns 31 km weather data into 100 m city temperature maps, a full year in 46 s on one GPU
- Five years shipping web and mobile apps in React, Node.js and Flutter
- Lately data engineering with Brussels open data: Azure, dbt, DuckDB, Terraform
- Open to AI, data and software roles in Brussels or elsewhere in Europe
- Arabic, French, English
| Project | What it is |
|---|---|
| Urban-climate emulator | MSc thesis. Residual U-Net emulating the UrbClim model for 100 European cities, RMSE 2.00 to 1.23 °C |
| Brussels Lakehouse · live | Azure lakehouse recording transit, air and grid data every 5 min, dbt + DuckDB |
| Brussels Live · live | Real-time map of every STIB vehicle, positions estimated without GPS |
| Image super-resolution (EDSR) | x4 super-resolution on DIV2K, 30.41 dB PSNR, +2.32 dB over bicubic |
| Weather forecasting with ML | Regression, random forests and neural networks on forecast and station data |
| CoAP over DTLS | Secure messaging for constrained IoT devices |
| iNeighbourHelper | Neighbourhood platform in Java with Dijkstra routing |
| notlightbloo | Image hosting: JWT auth, REST API, MySQL, React |
Python, PyTorch, xarray, scikit-learn · JavaScript, TypeScript, React, Node.js, Flutter, Dart, Java · Azure, dbt, DuckDB, Terraform, Docker, PostgreSQL, MySQL, Linux, Git



