Alaeddine Yacoub

Technical Product Owner
Jan 2024 – Present

• Coordinate technical delivery across 20+ engineers at BMW and two external development partners in three countries, translating stakeholder needs into technical requirements, delivery priorities and a shared roadmap.
• Manage a €1M+ budget for external development partners, overseeing financial planning, procurement and budget burn-down tracking.
• Rolled out a semantic search portal for fleet driving data, cutting scenario discovery time from days to minutes and helping teams curate safety-relevant and long-tail cases for model retraining and validation.
• Co-organized a cross-departmental series of hands-on AI workshops and hackathons for 200+ employees on building and applying agentic AI workflows.

Autonomous Driving Software Engineer
Sep 2022 – Dec 2023

• Developed core components of a data enrichment engine, making 100,000+ hours of driving data searchable through HD-map labels, signal-grounded VLM annotations and image-text embeddings to enable scenario discovery, similarity search and targeted data curation.
• Productionized an AWS-based ground-truth pipeline, using independent roof-mounted LiDAR and cameras to generate object, lane and road-boundary labels for benchmarking autonomous driving perception models.

Intern / Master’s Degree Candidate
Apr 2021 – Aug 2022

• Researched the effects of simulated point cloud distortion on real-time 3D object detection models across diverse driving scenarios.
• Developed calibration and distortion-correction algorithms for solid-state LiDAR point clouds distorted by cover surfaces.

• Applied signal processing methods (wavelet and fast Fourier transforms) to remove motion artifacts from PPG signals, improving respiratory-rate prediction accuracy by 15% on average.

• Migrated DeepVision Platform models (an in-house anomaly detection suite) to TensorFlow 2.0, enabling eager execution, improving flexibility and reducing debugging time by over 50%.

• Designed and trained a bidirectional LSTM model to forecast the short-term power output of 22 photovoltaic plants across Karlsruhe, improving accuracy by 13% over the chair’s SARIMAX baseline.
• Published the results at the 55th International Universities Power Engineering Conference (UPEC 2020).

• Provided on-site third-level IT and network support for 8,500+ employees worldwide.