Dr.-Ing. Ahmad Goudah

EU EFRE-funded AutoBin project with DST, SRS, and HGK Shipping across a 105 m inland vessel, an instrumented road vehicle, and driving simulation.

• Resolved buoy misclassifications and 5 missing waterway classes across 6 YOLO baselines by training one 85-class YOLOv8x model for road and water in PyTorch, reaching 0.960 person AP, 0.857 crane validation mAP@0.5, and 3.3 ms A100 inference per 640x640 image.
• Addressed camera darkness and glare limits by projecting 1,000 3D LiDAR scans into height-encoded maps with Open3D for a parallel 6-class YOLOv8 detector and 8-cluster K-Means ranging, localizing buoys, bridges, and shorelines from roughly 38 m to 285 m.
• Unified disconnected trial logs recorded by the AutoBin team on a 105 m vessel by fusing 6 cameras, 2 LiDARs, 20 Hz RTK DGPS, and AIS into a Multi-Sensor Fusion web tool with OpenCV stitching, diagnosing glare and hull occlusions.
• Built a 300 ms road-vehicle sensor platform pairing ESP32 and Raspberry Pi 4 over Modbus RTU to log LiDAR, 4 GPS, 5 ultrasonic, and IMU streams for Unity 3D replay, and applied a CNN+LSTM pre-collision risk approach in CARLA simulation.

University software and data engineering unit; owned institutional platforms and data systems (remote advisory during German doctorate, 2019-2025).

• Led a team of 4 engineers automating university workflows across 4 core platforms (admissions, payroll, elections, surveys) and the publishing portal (apc.aast.edu) in C#, ASP.NET Core, SQL Server, and CI/CD, serving thousands of students and staff in production.
• Replaced manual spreadsheet reporting across 3 institutional domains (HR, purchasing, and student outcomes) by building interactive Business Intelligence KPI dashboards and tuning SQL Server queries and REST APIs, cutting recurring executive report generation from days to minutes.
• Addressed smart-grid phase unbalance and demand uncertainty by developing Machine Learning models for 24-hour active power forecasting and Genetic Algorithms for three-phase load balancing in Python and MATLAB, validating results across 4 peer-reviewed energy studies.

• Supervised 10+ automated driving and robotics theses to completion and taught Python and C++ computer vision and embedded labs to hundreds of students across 3 universities.

• Founded Al-Amal to design and ship commercial AVR Microcontroller LED displays in C/C++, and built automated cheque, archiving, and access-control software for Toyota ALJ and Noortech.