I am a final year DPhil (PhD) student at MRG and CRG at the Oxford Robotics Institute, University of Oxford, supervised by Daniele De Martini, Paul Newman, and Lars Kunze. For my studies, I have been awarded the Oxford - Google DeepMind research scholarship.
Currently, I am also working part-time at OXA as Senior ML Engineer, at the Office of the CTO. Before starting my DPhil, I was working as Lead Self-Driving Software Engineer at StreetDrone (acquired by OXA) and worked as Data Engineer at Williams Racing Formula 1.
I finished my MEng in 2017 at NTUA, Greece, Department of Electrical and Computer Engineering. During my studies, I spent time at ETH Zürich conducting my master’s thesis under the supervision of Prof Luc Van Gool and Dr Dengxin Dai, working with AMZ Driverless.
My research focuses on introspective robot learning and neurosymbolic AI, with work on graph learning, scene understanding, neural algorithmic reasoning, semantic localisation, and explainability.
Latest Research
NAR-* ICP: Neural Execution of Classical ICP-based Pointcloud Registration Algorithms
Efimia Panagiotaki, Daniele De Martini, Lars Kunze, Paul Newman, and Petar Veličković
Neural algorithmic reasoning for point cloud registration, emulating the reasoning steps of classical algorithms.
Tackling GNARLy Problems: Graph Neural Algorithmic Reasoning Reimagined through Reinforcement Learning
Alex Schutz, Victor-Alexandru Darvariu, Efimia Panagiotaki, Bruno Lacerda, and Nick Hawes
A framework that frames algorithmic trajectory learning as an MDP and applies imitation learning and RL to solve combinatorial NP-hard problems.
GraphSCENE: On-Demand Critical Scenario Generation for Autonomous Vehicles in SimulationIROS'25
Efimia Panagiotaki, Georgi Pramatarov, Lars Kunze, and Daniele De Martini
A temporal GNN-based method to automatically generate diverse, safety-critical traffic scenarios in simulation on-demand, grounded on real-world datasets.
The Oxford RobotCycle Project: A Multimodal Urban Cycling Dataset for Assessing the Safety of Vulnerable Road UsersT-FR'25
Efimia Panagiotaki, Divya Thuremella, Jumana Baghabrah, Samuel Sze, Lanke Frank Tarimo Fu, Benjamin Hardin, Tyler Reinmund, Tobit Flatscher, Daniel Marques, Chris Prahacs, Lars Kunze, and Daniele De Martini
Large-scale multimodal dataset, including eye-gaze data, capturing first-person ego-cyclist perspective.
OORD: The Oxford Offroad Radar DatasetT-ITS'24
Matthew Gadd, Daniele De Martini, Oliver Bartlett, Paul Murcutt, Matt Towlson, Matthew Widojo, Valentina Muşat, Luke Robinson, Efimia Panagiotaki, Georgi Pramatarov, Marc Alexander Kühn, Letizia Marchegiani, Paul Newman, and Lars Kunze.
Radar and GPS/INS Dataset collected in the rugged Scottish highlands in extreme weather