Aleksandr Kazakov

• Designed and developed an intelligent document processing (IDP) system for robust and comprehensive OCR and PDF parsing • Started as the sole developer and progressed to leading a small team • Applied a wide range of techniques, from classical computer vision and algorithmic approaches to deep learning and modern multimodal LLMs • Maintained the integration of the system into the company’s API and desktop distribution

• Conducted research on wave process simulation in rigid bodies with complex 3D shapes and rheology • Proposed a novel version of grid‑characteristic numerical method for non‑periodical tetrahedral grids • Implemented the method in a scalable high‑performance C++ application using MPI • Successfully applied the simulation to diverse tasks, from medical ultrasound - https://doi.org/10.1515/rnam-2016-0030 - to fiber composites delamination - https://doi.org/10.1134/S1995080218070065

PhD thesis: "Numerical simulation of wave processes in ultrasonic non-destructive testing by the grid-characteristic method". Developed and applied a version of the grid-characteristic method for non-periodic tetrahedral grids for complex geometry regions of integration