Sebastian-Ion Nae

PhD Student, ACS · University Politehnica of Bucharest

Machine Learning Research Engineer @ Heimdal Security

I'm a PhD student at ACS, University Politehnica of Bucharest, working on efficient learning from images and video under minimal supervision and on the safety of AI systems. My PhD is officially titled "Distributed Task-Oriented Navigation in UAV Swarms through Multimodal Scene Understanding," but the real question I'm chasing is broader: how systems can learn efficiently from limited, streaming visual data, with drones as the embodied agents generating those observations.

Alongside the PhD, I work at Heimdal Security on anomaly detection, AI safety systems, and patch prioritization using unsupervised methods.

I occasionally write on Substack about side projects and experiments worth sharing, an experiment in itself, to get better at writing.

Publications

See Google Scholar for the full list of publications.

IndoorCrowd: A Multi-Scene Dataset for Human Detection, Segmentation, and Tracking with an Automated Annotation Pipeline Sebastian-Ion Nae, Radu Moldoveanu, Alexandra Ștefania Ghiță, Adina Magda Florea CVPR Workshops (DataCV) · 2026 CVF · arXiv · Project Page · Dataset
Learning on the Fly: Replay-Based Continual Object Perception for Indoor Drones Sebastian-Ion Nae, Mihai-Eugen Barbu, Sebastian Mocanu, Marius Leordeanu European Robotics Forum (ERF) · 2026 arXiv · Project Page
Efficient Self-Supervised Neuro-Analytic Visual Servoing for Real-time Quadrotor Control Sebastian Mocanu, Sebastian-Ion Nae, Mihai-Eugen Barbu, Marius Leordeanu ICCV Workshops (Embedded Vision Workshop, Oral) · 2025 CVF · arXiv · Code · Project Page

Experience

See LinkedIn for the full work history.

Machine Learning / AI Research Engineer, Heimdal Security 2023 – present

Anomaly detection, AI safety systems, and patch prioritization using unsupervised methods.
Data Scientist, BCR Nov 2022 – Dec 2023

Anomaly and fraud detection models, reporting pipelines, and training dataset creation (Python, Scikit-learn, KNIME, Spark, Hive, SQL).
Research Engineer Intern, Expleo Jun 2022 – Oct 2022

CNN-based image recognition, PID control software for a custom drone, and bare-metal embedded development (TensorFlow, Python, C/C++).

Education

PhD, ACS · University Politehnica of Bucharest 2025 – present

Thesis: Distributed Task-Oriented Navigation in UAV Swarms through Multimodal Scene Understanding (supervisor: Adina Magda Florea).
MSc in Artificial Intelligence, ACS · University Politehnica of Bucharest 2023 – 2025

Thesis: Efficient Self-flying Visual Drones for Object Following, a GPS-free indoor drone tracking system combining vision-language models with lightweight YOLOv8 detection for real-time 2D visual control (supervisor: Marius Leordeanu).
BEng in Computer Engineering, FILS · University Politehnica of Bucharest 2019 – 2023

Thesis: MalwareNET: Classification of Malware Files Using Convolutional Neural Networks, which converts malware files into color images for CNN-based classification, achieving high F1 scores on the Microsoft BIG 2015 dataset (supervisor: Iulia Stanica).