Dr. Viktor Begun

- LLM agent system: Orchestration of the Claude API for autonomous CV/cover letter generation. Implementation of anti-hallucination, context engineering, and 100% reproducibility. - Azure-hosted chatbot optimization: achieved 5x faster response with full RAG context. - Developed and productionized chatbot voice extension: Text to Speech and Speech to Text, including GPT-Realtime, and demo site. - Dockerized and deployed chatbot on Azure VM with NSG and end-to-end latency 2-3s.

- Translated business requirements into technical tasks and created interactive PowerBI dashboards to support management decisions. - Analyzed manufacturing data: camera and sensors, to monitor production quality. - Built and productionized advanced time series analysis algorithms using Python, to automatically flag defects, reduce manual checks, and ensure stable, scalable quality assessment pipelines.

- Developed a 30% faster algorithm to detect and categorize processing errors in a large-scale document recognition and conversion pipeline, including Word, OpenOffice, and PDF documents. - Designed a priority system to highlight the most critical failures, helping the team focus on the errors with the most significant business impact.

- Created long-term turbine-health forecasts for an extremely imbalanced data ( >40 M train- and <5 test-points ) to predict failures and allow early interventions. - Successfully applied a combination of machine learning, probability theory, and quality control methods to find the most cost-effective management strategy. - The average savings obtained for the hydraulic group have the amount of 85 200 Euro, and the maximum of 173 100 Euro per 5 considered turbines per year.

- Practicum in physics laboratory: teaching, managing, evaluation (in German). - Mathematics exercises: teaching, evaluation of exams (in German). - Preparation of the content materials for the “Big Data Analytics” course (Python / Pandas).

- Teaching 7 different physics courses, including Nuclear Physics and Quantum mechanics (in Polish and English) - Developed a course of lectures on General Physics with Exercises from scratch - 6 Publications in international physics journals - Referee for several physics journals, including Physical Review C from American Physical Society - Coordinator of the online seminars “NA61-theory” https://indico.cern.ch/category/5919/ - Habilitation

- 16 Publications, including 10 journal publications, 6 conference proceedings (in English). - Statistical thermal model analyses of particle production in heavy-ion collisions using SHARE program (in Fortran) and THERMUS program (in C++) - Analysis of particle spectra in heavy ion collisions using ROOT based THERMINATOR program - Peer-Review of a candidate from University of Cape Town for South Africa’s National Research Foundation

• Editor of the book “Progress in High Energy Physics and Nuclear Safety (NATO Science for Peace and Security)” • Correction and systematization of the authors contributions in order to make consistent, accurate, and complete proceedings of the conference • Logistic and scientific organization of international conferences, Ukraine • Taking care of arrival, departure, hotel reservation, translation, solving problems of foreign guests of the conference • Planning and update of the conference schedu

• DFG Cooperative Grant, participant, Germany (02 – 08/2012) • Numerical modeling of the particle distributions in heavy ion collisions using HSD transport model (in Fortran) for NA61 experiment at CERN • The results are added to NA61 database at CERN and are used for comparison with experimental data • Statistical thermal model analyses of particle production in heavy-ion collisions using THERMUS, a package of C++ classes and functions

• Several analytical models of phase transitions in heavy ion collision were studied and improved • Consulting of experimental group in Joint Institute for Nuclear Research, Russia • Feasibility studies, definition and optimization of experimental conditions for the THERMALISATION group that search for the anomalous fluctuations predicted by our model • Peer-Review of papers proposed for publication in nuclear physics section of International Journal of Modern Physics E

• INTAS Young Scientist Competitive Fellowship, from European Union (2007 – 2008) • Previously developed model for calculations of particle fluctuations was improved by including all effects relevant to experiment

• INTAS Young Scientist Competitive Fellowship, from European Union (2007 – 2008) • Experimental observation of the fluctuation effect predicted by our model http://inspirehep.net/record/732305?ln=en

• Guest Scientist in University of Florence, Italy • The observed threshold effect for average number of particles was tested numerically in close collaboration with scientists from Florence

PhD Thesis: “Fluctuations and Conservation Laws in Relativistic Hadron Gas”