Ana-Maria Marcu
Welcome to my web page! I’m Ana-Maria, an Applied Scientist at Wayve with an MEng in Electrical and Information Sciences from the University of Cambridge. My research centers on representation learning for embodied intelligence. Currently, I focus on developing challenging learning objectives for end-to-end neural networks, with the goal of advancing reliable autonomous systems. I have pioneered one of the first open-vocabulary evaluation frameworks for visual question answering in autonomous driving. This work led to the creation of the Lingo model family, which has been featured in MIT Technology Review and the Financial Times, and published at ECCV and presented at CVPR. My mission is to benefit humanity through the safe development of artificial intelligence.
Community
Embodied Intelligence for Autonomous Systems on the Horizon [CVPR 2025]
A workshop discusses the present and future of autonomous systems - what progress has been made and what challenges still remain.
WorkshopWomen in Computer Vision [CVPR 2024, ECCV 2024, CVPR 2025]
I am passionate about supporting women towards fullfilling their potential in an AI research career.
WebsitePublications
LingoQA - Visual Question Answering for Autonomous Driving [ECCV 2024]
LingoQA is a novel dataset and benchmark for visual question answering in autonomous driving.
PaperBlog Posts
Lingo-2 - Driving with Natural Language
Lingo-2 is a driving model that links vision, language, and action.
BlogLingo - Natural Language for Driving
Lingo is a vision-language model that takes a short video as an input and outputs explanations in response to questions related to autonomous driving.
BlogLoupe 360 - Automated Tunnel Inspection
Loupe 360 is a tunnel inspection visualisation and analytics platform. It includes object detection models that identify defects such as cracks and water ingress.
Blog