Tarek Bouamer

Hi, I’m Tarek

I’m a Research Staff Member at Layer42, working on spatial AI, 3D computer vision, and robotics. My work connects research with practical systems that perceive, reconstruct, and reason about the physical world.

Research interests

• Visual localization: Image retrieval, place recognition, and image matching using learned features and foundation models.
• 3D reconstruction and mapping: Structure from motion, SLAM, feed-forward reconstruction, and Gaussian Splatting.
• 3D world models: Compact, scalable 3D state representations for navigation and embodied intelligence.
• Multimodal spatial reasoning: Connecting vision, geometry, and language for 3D understanding and language-guided interaction.
• Robot learning: Vision-language-action (VLA) models, imitation and reinforcement learning, and model predictive control for navigation and manipulation.
• Earth observation: Representation learning, EO foundation models, and efficient VLMs for satellite and aerial imagery.

Writing and open source

I share technical articles, research implementations, and open-source tools. Examples include the 3D-LLM series, exploring language models for 3D understanding, and EO-VLM, a benchmarking framework for vision-language models on Earth observation tasks.

• Projects: Research implementations and tools.
• Blog: Technical tutorials and research reviews.
• CV: Experience, education, and technical background.

For research opportunities, collaboration, or R&D projects through Layer42, visit Contact.