Hungary
Data Scientist with 5+ years of experience applying Machine Learning and Deep Learning in industrial and engineering contexts, with a strong foundation in Mechanical Engineering. I have delivered solutions ranging from predictive maintenance models for pumps and image processing algorithms for semiconductor equipment, to combinatorial optimization algorithms for logistics, LLM/RAG applications for production analysis, and frontend–backend features for industrial IoT platforms. I have worked independently on projects from concept to deployment, collaborated in cross-functional teams, and led AI development efforts. Through my education in Mechanical Engineering and hands-on work in Data Science, I have developed the ability to understand machines and physical processes in depth and to apply data-driven methods in a way that is both optimal and practical. My mission is to keep growing as a specialist by taking part in impactful projects, building robust AI solutions, and bridging the gap between physical engineering knowledge and modern data science.
▪ Leading end-to-end data science consulting engagements, including scoping, high-level solution architecture, task breakdown, and delivery planning. ▪ Delivering ML/AI solutions in mixed DS/DE teams while acting as the client-facing technical lead / SPOC across EMEA and US customers. ▪ Advising clients on best-practice implementations leveraging Databricks and cloud-native capabilities (Azure/AWS), and running enablement workshops. ▪ Developing and operationalizing ML/LLM use cases (e.g., predictive modeling, NLP/RAG) with strong software engineering practices (clean code, git, testing). ▪ Contributing to internal solution development and research initiatives, and supporting pre-sales, courseware, and knowledge-sharing within the DS guild.
▪ Developed features for INNIO's custom IoT platform (myplant), combining frontend dashboards and backend data integrations. ▪ Built highly interactive frontend analysis dashboards for engineers based on Dash and Plotly, including a dynamic graph definition tool, enabling better data exploration and visualization of industrial equipment data. ▪ Implemented backend-to-database integrations, ensuring smooth data retrieval and processing from ClickHouse and PostgreSQL for analytics workflows. ▪ Collaborated with global teams (US, Austria) in a cross-functional setting, bridging communication with engineers and guiding analysis decisions for ambiguous/incomplete data.
▪ Delivered data-driven solutions for semiconductor production by analyzing workflows and identifying opportunities where AI could be applied. ▪ Performed feasibility studies with data scientists and domain experts, secured management approval, and managed AI development projects. ▪ Designed and trained ML/DL models covering predictive maintenance, image processing, root cause analysis, natural language processing, and RAG. ▪ Developed a combinatorial optimization solution for vehicle route planning, improving logistics efficiency. ▪ Supported the integration of AI models into production systems, ensuring robustness and alignment with manufacturing requirements.
▪ Developed a predictive maintenance model from scratch in Python for a series of vacuum pumps, enabling early detection of failures. ▪ Achieved potential failure detection up to one month in advance, increasing reliability of maintenance planning. ▪ Visualized tens of millions of sensor data points through interactive dashboards, improving engineers’ ability to explore and interpret machine behavior. ▪ Collaborated with domain experts to incorporate knowledge of machine physics, operational modes, and potential failure types into the modeling process. ▪ Deployed both traditional ML models and LSTM-based deep learning models in parallel, benchmarking their performance for predictive accuracy.
▪ Built a DNN regression model to aid fluid dynamic simulations around 3D printed parts. ▪ Estimated the frictional effects of surface textures on the surrounding flow using AI models. ▪ Increased simulation speed and accuracy, reducing computational cost while maintaining reliability.
Contributing to a research project about modelling turbulent atmospheric flows by building a Deep Learning model to create a special turbulence related boundary condition. ▪ Writing an academical summary of the theory behind DNN’s, RNN’s, and LSTM’s complete with intuitive explanations and equations. ▪ Developing an LSTM-based model that can be incorporated in fluid dynamic simulations to model the effect of turbulence.