Internship Programme at GATE Institutе

Worked on integrating large language models with physical task execution in robotics. His focus was on embodied AI and Vision–Language–Action models, including dataset processing, experimental design, and model evaluation. This work contributed to a paper accepted at ACM SAC 2026 and to the release of the GATE-VLAP dataset.

Processed and annotated textual datasets for training machine learning models detecting disinformation and automated text. Contributed to a social media analysis tool and annotated datasets. Co-author of a scientific paper presented at the 10th Language & Technologies Conference (Poznań, 2023).

Worked on social media disinformation analysis, including large-scale data preprocessing and development of analytical tools. Contributed to Bulgarian-language datasets and a scientific publication

Contributed to analytical and administrative tasks related to establishing a Big Data and AI test centre within NATO DIANA.

Worked on information synthesis, structuring and media coverage analysis related to the project.

Collected, filtered, visualised and analysed data on urban heat islands. Contributed to a scientific publication.

Collected and annotated data for disinformation and text deepfakes, tested and improved models, developed Python scripts.

Developed a model for propaganda detection under limited-resource conditions and created new datasets.

Improved AI models, tested web tools, analysed model errors and collected new data.

Optimised walkability index calculations and developed a graphical environment for scenario testing.

Developed an automated system for detecting tree canopies using computer vision, machine learning and remote sensing.