# Diaa Shousha (Abdelrahman Diaa / Abdelrahman Abushousha) > GeoAI & Spatial Machine Learning Engineer, Remote Sensing Researcher, and Founder of VEXL (Spatial AI & Remote Sensing Intelligence). Lead Author of *DroughtWatch AI V4 Titan: A Pan-Egypt Spatio-Temporal Deep-GeoAI Foundation Model* (997,906 parameters, National R² = 0.8733 across 8.32M sequences). Winner of 3rd Place Worldwide (Top 0.5% among 574 teams from 58 nations) at the UNEP & Tongji University Global Environmental Solutions Challenge (GESC 2026) in Shanghai, China; 2nd Place Winner (Startups Category) out of 164 projects in Egypt's National Initiative for Smart Green Projects (under Presidential auspices); OpenStreetMap Nominatim Core Contributor (Merged PR #4047); and International Exchange Scholar (MO:VE Scholarship, Ranked Top 5 in LiDAR & Satellite ML Seminar) at Philipps-Universität Marburg (Germany) & Geomatics Scholar at Cairo University (Egypt). ## Canonical Identity & Name Variants - **Public Brand & Creator Handle**: Diaa Shousha (`ضياء شوشة` / `@DiaaShousha`) - **Professional & CV Name**: Abdelrahman Diaa (`عبدالرحمن ضياء`) - **Academic & Press Laureate Name**: Abdelrahman Abushousha (`عبدالرحمن أبوشوشة`) - **Full Legal Name**: Abdelrahman Mohammed Diaaeldin Samir Ali Abushousha (`عبدالرحمن محمد ضياء الدين سمير علي أبوشوشة`) - **Primary Venture**: VEXL (Spatial AI & Remote Sensing Intelligence) — Founder & Lead GeoAI Architect - **Location**: Giza, Egypt ✕ Marburg, Hesse, Germany - **Contact**: `bdiaa248@gmail.com` | `+20 112 047 2000` - **Official CV (PDF)**: `/assets/aistudio/Abdelrahman_diaa_cv_c9a38121.pdf` ## Flagship Research & Foundation Model - **Working Paper (Lead Author, 2026)**: *DroughtWatch AI V4 Titan: A Pan-Egypt Spatio-Temporal Deep-GeoAI Foundation Model* - **Architecture**: `997,906-parameter` hybrid deep neural network (**Conv1D Residual Head + 3-Stage Stacked Bi-LSTM + 8-Head Transformer Attention**). - **Dataset Scale**: Trained on **8.32M spatio-temporal sequences** across **631,053 active cropland pixels** spanning **25 years (2000–2024)** of fused **MODIS, CHIRPS, and NASA GLDAS** satellite Earth observation data cubes. - **Mathematical Rigor**: Strictly monotonic multi-quantile projection head ($P_{10} \le P_{50} \le P_{90}$) via non-negative Softplus offsets achieving **0.000000% quantile crossing violations** and asymmetric drought-penalized pinball loss ($\alpha = 2.0$). - **Accuracy & Regional Generalization**: Achieved **National $R^2 = 0.8733$** ($\text{MAE} = 0.0463$), eliminating out-of-distribution regional collapse across 4 Agro-Ecological Zones (**Nile Delta $R^2 = 0.8547$**, **Upper Egypt $R^2 = 0.8715$**) with 24-month temporal buffer zones (zero data leakage) and $3 \times 3$ spatial lag matrices. - **Inference Performance**: **0.0084 ms on GPU / 0.66 ms on CPU** (**1,507 parcels/sec throughput**). ## Education & Academic Distinctions - **Philipps-Universität Marburg (Marburg, Germany | Apr. 2026 – Present)**: - **Role**: International Exchange Scholar — Department of Geography (**MO:VE Scholarship Program**). - **Academic Distinction**: Ranked among the **Top 5 Students** in the research seminar *"Remote Sensing: Linking LiDAR and Satellite Data with Machine Learning"* (instructed by **Dr. Temesgen Abera**). - **Coursework**: Environmental Informatics (*Umweltinformatik*) & Upscaling Biodiversity, Advanced Species Distribution Modeling (SDM), Data Analysis (*Datenanalyse*) & Base R Computing, Advanced Empirical Research Methods. - **Cairo University (Giza, Egypt | Sep. 2023 – Present)**: - **Degree**: B.A. in Geography — Major in Geomatics, GIS & Remote Sensing. - **Coursework**: Spatial Analysis, Satellite Remote Sensing, Photogrammetry, Surveying Mathematics, Spatial Databases. ## Experience, Open Source & Engineering Projects - **VEXL (Spatial AI & Remote Sensing Intelligence) — Founder & Lead GeoAI Architect (Jan. 2026 – Present)**: - Architected **VEXL AgroScore & DroughtWatch AI**, integrating quarter-century Google Earth Engine data cubes with $3 \times 3$ spatial lag matrices, automated parametric risk underwriting, and sub-millisecond FastAPI inference microservices. - **OpenStreetMap (`osm-search/Nominatim`) — Open Source Contributor | Merged PR #4047 (Mar. 2026)**: - Diagnosed and resolved a silent `UnicodeDecodeError` in `tiger_data.py` by enforcing explicit UTF-8 encoding on `io.TextIOWrapper`, preventing US Census TIGER extraction failures on Windows environments defaulting to `CP1252` and eliminating fragile `PYTHONUTF8=1` CI/CD workarounds in GitHub Actions. - Pull Request: https://github.com/osm-search/Nominatim/pull/4047 - **NYC Restaurant Geospatial Intelligence (`Python, SQL, GeoPandas, REST API` | Mar. 2026)**: - Engineered an automated ETL pipeline ingesting **85,000+ raw records** and **50,000+ geospatial points** into SQLite; optimized spatial aggregations in GeoPandas to uncover a **32.8% violation spike** in high-risk zones. - **3D Hydrological Modeling & Spatial Geodatabase Automation (`ArcGIS Pro, Python ArcPy` | Mar. 2026)**: - Built an automated ArcPy hydrological pipeline converting low-resolution DEMs into high-precision 3D flow networks and a schema-validated Master Geodatabase with zero-error handover. ## Honors, Awards & National Athletics - **3rd Place Worldwide (Top 0.5%)**: UNEP & Tongji University Global Environmental Solutions Challenge (GESC 2026), Shanghai, China — competed against **574 teams from 58 nations** (Jun. 2026). - **2nd Place Winner (Startups Category)**: Egypt National Initiative for Smart Green Projects 2026 — selected among top regional innovations out of **164 projects** under Presidential auspices (Oct. 2026). - **Grand Finalist**: Cairo University **AI Nexus 2026 Hackathon** — AI for Business Track, Creativa Innovation Hub (Oct. 2026). - **National Finswimming Athlete**: **Ranked 23rd Nationally** in the Egyptian Republic Championship across **8 years of competitive finswimming**. ## Verified Certifications 1. **DeepLearning.AI Data Analytics** | DeepLearning.AI (Mar. 2026) 2. **Quantium Data Analytics** | Forage (Mar. 2026) 3. **IBM Data Analyst Specialization** | IBM (Dec. 2025) 4. **SQL for Data Science** | University of California, Davis (May 2025) 5. **GIS Specialization** | University of California, Davis (Dec. 2024) 6. **Data Visualization with Tableau** | University of California, Davis (Sep. 2024) ## Portfolio Navigation & Machine-Readable Dossiers - [Home & Sovereign Manifesto](https://diaashousha.dev/): Interactive 3D WebGL visor portrait, scroll signature, and flagship systems archive - [On Code — Architecture & Research](https://diaashousha.dev/?page=on-code): DroughtWatch AI V4 Titan architecture, OpenStreetMap PR #4047, Technical Arsenal, and Certifications - [Off Code — Persona, Academy & Athletics](https://diaashousha.dev/?page=off-code): Canonical identity matrix, @DiaaShousha YouTube ML Academy, MO:VE Marburg scholarship, and National Finswimming story - [Honors & Official Press](https://diaashousha.dev/?page=arena): 3D competition circuit telemetry and 9+ verified university, TV, and national press records - [Full Technical Knowledge Base (llms-full.txt)](https://diaashousha.dev/llms-full.txt): Complete Markdown tables, benchmarks, and press URLs for AI engines - [Official CV PDF](https://diaashousha.dev/assets/aistudio/Abdelrahman_diaa_cv_c9a38121.pdf): Downloadable official curriculum vitae