Vojtěch Sýkora

• Independently create a comprehensive dataset of Raspberry Pi schematics for computer vision analysis. • Refine a multi-modal pipeline to extract semantics from schematics, benchmarking it against state-of-the-art Visual Language Models (VLMs). • Engineer end-to-end components including object detection, instance segmentation, rotation recognition, text recognition, and edge detection to improve information ex- traction and data quality.

• Lead the development of an agentic LLM+VLM prototype for advanced data processing. • Adapt and optimize an existing multi-modal pipeline to extract semantics and graph structure from switchboard schematics. • Identify and classify visual components in complex electrical diagrams to support down- stream automation for enterprise clients. • Implement targeted refinements to improve pipeline robustness and accuracy.

• Conducted research on identifying the most congested areas in urban environments, fo- cusing on Dublin and Luxembourg. • Developed and adapted gravitational clustering techniques to detect high-traffic regions, optimizing urban mobility analysis. • Worked with publicly available traffic datasets, applying data-driven approaches to iden- tify problematic traffic zones. • Co-authored a paper submitted to Expert Systems with Applications (ESWA), Elsevier.

• Project 1 (2020): Analyzed 600K+ records to gain insights on internal migration flows in the Czech Republic. (MySQL, Python) • Project 2 (2022): Developed an animated choropleth map of geographic distribution of allowance for living (Python, GeoJSON, NumPy, Pandas, Plotly).

• Contributed to the FRAS (Flexible and Resilient Autonomous Systems) research project, funded by the US Air Force Research Lab. • Developed a Python and PDDL-based environment generator to train single and multi- agent AI planning strategies. • Applied classical planning and game theory concepts to create robust algorithms for adversarial environments.

• Full scholarship from the DAAD German Academic Exchange Service for whole study program of 2 years. • Focused on Deep Learning, Computer Vision, and Reinforcement Learning. • Thesis: Multi-modal Deep Learning for Automated Schematic Analysis

• Computer Science studies with a focus on Artificial Intelligence. • Thesis: Proximal Policy Optimization for Car Racing with unpredictable Wind. • Data Structures & Algorithms, C/C++, Java, Python, Parallel Computing